1 RMU72 Oracle RMU, the Oracle Rdb management utility, lets database administrators manage Oracle Rdb databases. Oracle RMU commands are executed at the operating system prompt. Oracle RMU command syntax follows the rules and conventions of the DIGITAL Command Language (DCL). See the RMUALTER help also. Oracle RMU commands allow you to display the contents of database files, control the Oracle Rdb monitor process, verify data structures, perform maintenance tasks (such as backup and restore operations), and list information about current database users and database activity statistics. Oracle RMU commands consist of words, generally verbs, that have parameters and qualifiers to define the action to be performed. 2 Command_Parameters One or more spaces separate command parameters and their qualifiers from the command keyword. Command parameters define the file, index, or entity on which the command will act. In most cases, you can omit the parameter from the command line and enter it in response to a prompt. In the following sample command, RMU/DUMP is the command keyword and MF_PERSONNEL is the command parameter: $ RMU/DUMP MF_PERSONNEL When a storage area is a command parameter in an Oracle RMU command, use the storage area name instead of the storage area file specification. For example: $ RMU/RESTORE/AREA MF_PERSONNEL.RBF EMPIDS_LOW/THRESHOLDS=(65,75,80) Some commands, such as the RMU Backup command, require two or more command parameters. If you provide all parameters (for example, a root file specification and a backup file name), there are no prompts. Other commands, such as the RMU Restore command, have one required and one optional command parameter. In this case, there are no prompts if you provide the backup parameter but not the storage area parameter. However, if you do not provide either parameter, Oracle RMU prompts for both. 2 Command_Qualifiers Command qualifiers modify the behavior of an Oracle RMU command. Although similar in appearance, command qualifiers are different from the Oracle RMU commands themselves. The first (and sometimes the subsequent) word that follows the RMU keyword is the command itself. For instance, in the following example, /DUMP and /AFTER_ JOURNAL are part of the Oracle RMU command and thus must appear in the order shown. /OPTION=STATISTICS and /LOG are command qualifiers and can appear in any order after the Oracle RMU command. You can determine which portions of an Oracle RMU command are the command itself, and which portions are command qualifiers by noting the documented name of the command, $ RMU/DUMP/AFTER_JOURNAL aij_one.aij /OPTION=STATISTICS/LOG Command qualifiers can be entered as upper-, lower-, or mixed- case type. They always begin with a slash (/) followed by a qualifier word. In some cases, an equal sign (=) and a qualifier value follow the qualifier word. A qualifier value can be simple (a number, a string, or a keyword) or compound (a list of numbers, strings, or keywords separated by commas, enclosed in parentheses) or an indirect command file name. For information on using indirect command files, see Indirect-Command-Files. A default value for a qualifier indicates what qualifier will be used if you omit the qualifier completely. Omitting a qualifier is not the same thing as specifying a qualifier with a default argument. Command qualifiers influence the overall action of a command. Command qualifiers must be placed following the command keyword but before any parameters. In the following example, the command qualifier, Users, immediately follows the Dump keyword and precedes the command parameter, mf_personnel: $ RMU/DUMP/USERS MF_PERSONNEL Parameter qualifiers (also referred to as file qualifiers or area qualifiers) affect the treatment of parameters in the command. If the command includes multiple instances of a given type of parameter, the placement of parameter qualifiers affects their scope of influence as follows: o If you position the parameter qualifier after a particular parameter, the qualifier affects only that parameter. This is local use of a parameter qualifier. o If you position the parameter qualifier before the first parameter, the qualifier applies to all instances of the parameter. This is global use of a parameter qualifier. Not all parameter qualifiers can be used globally. To identify such qualifiers, read the description of the qualifier. o If you position the parameter qualifier after a parameter, the qualifier applies only to that instance of the parameter. Local parameter qualifiers take precedence over global parameter qualifiers, in most cases. Exceptions are documented in the qualifier descriptions for each Oracle RMU command. The following example demonstrates the local use of the area qualifier, Thresholds, to change the threshold settings for the EMPIDS_LOW area: $ RMU/RESTORE MF_PERSONNEL EMPIDS_LOW/THRESHOLDS=(70,80,90) Note that if you specify a qualifier in both the negative and positive forms, the last occurrence of the qualifier is the one that takes effect. For example, the Nolog qualifier takes effect in this command: $ RMU/BACKUP/LOG/NOLOG MF_PERSONNEL MF_PERS This is consistent with DCL behavior for negative and positive qualifiers. 2 Indirect-Command-Files Numerous Oracle RMU command operations accept lists of names as values for certain qualifiers, such as the Areas= or Lareas= qualifiers. The command syntax can easily exceed the maximum length of 1024 characters accepted by DCL. To overcome the problem of syntax that is too long, you can include the names in an indirect command file and specify the indirect command file following the qualifier. Throughout this manual, this is commonly referred to as using an indirect file reference. Note that indirect command files can be nested to any depth. Each indirect command file (default file extension .opt) contains a list of names with one name per line. A comment, preceded by an exclamation point, can be appended to a name, or it can be inserted between lines. A reference to an indirect command file in the list must be preceded by an at sign (@) and enclosed in quotation marks (""). For example: "@EMPIDS". The following example shows the contents of an indirect command file called empids.opt. It lists the EMPIDS_LOW, EMPIDS_MID, and EMPIDS_OVER storage areas. The last line in the example shows how you would reference the indirect command file in an Oracle RMU command line with the required quotation marks. $ TYPE EMPIDS.OPT EMPIDS_LOW ! Employee Areas EMPIDS_MID EMPIDS_OVER $ RMU/ANALYZE/AREA="@EMPIDS" MF_PERSONNEL ! ANALYZE EMPLOYEE AREAS 2 Required_Privileges An access control list (ACL) is created by default on the root file of each Oracle Rdb database. To be able to use a particular Oracle RMU command for the database, you must be granted the appropriate Oracle RMU privilege for that command in the database's root file ACL. For some Oracle RMU commands, you must have one or more OpenVMS privileges as well as the appropriate Oracle RMU privilege to be able to use the command. Note that the root file ACL created by default on each Oracle Rdb database controls only your Oracle RMU access to the database (by specifying privileges that will allow a user or group of users access to specific Oracle RMU commands). Root file ACLs do not control your access to the database with SQL (structured query language) statements. See Show Privilege for information on how to display your Oracle RMU access to the database. Your access to a database with SQL statements is governed by the privileges granted to you in the database ACL (the ACL that is displayed when you use the SQL SHOW PROTECTION ON DATABASE command). Privileges Required for Oracle RMU Commands shows the Oracle RMU privileges you must have to use each Oracle RMU command. When more than one Oracle RMU privilege appears in the Required Oracle RMU Privileges column, if you have any of the listed Oracle RMU privileges, you will pass the Oracle RMU privilege check for the specified Oracle RMU command. If the Oracle RMU command requires a user to have one or more OpenVMS privileges in addition to the appropriate Oracle RMU privileges, the OpenVMS privileges are shown in the Required OpenVMS Privileges column of Privileges Required for Oracle RMU Commands. When more than one OpenVMS privilege is listed in the Required OpenVMS Privileges column, you must have all of the listed OpenVMS privileges to pass the OpenVMS privilege check for the Oracle RMU command. The OpenVMS Override Privileges column of Privileges Required for Oracle RMU Commands shows one or more OpenVMS privileges that allow a user without the appropriate required Oracle RMU and OpenVMS privileges for an Oracle RMU command to use the command anyway. When more than one OpenVMS privilege is listed in the OpenVMS Override Privileges column, you can use the specified Oracle RMU command if you have any of the listed privileges. Table 1 Privileges Required for Oracle RMU Commands Required Oracle Required Oracle RMU RMU OpenVMS OpenVMS Override Command Privileges Privileges Privileges Alter RMU$ALTER SYSPRV, BYPASS Analyze Areas RMU$ANALYZE SYSPRV, BYPASS Analyze RMU$ANALYZE SYSPRV, BYPASS Cardinality Analyze Indexes RMU$ANALYZE SYSPRV, BYPASS Analyze RMU$ANALYZE SYSPRV, BYPASS Placement Backup RMU$BACKUP SYSPRV, BYPASS Backup After_ RMU$BACKUP SYSPRV, BYPASS Journal Backup Plan RMU$BACKUP SYSPRV, BYPASS Checkpoint RMU$BACKUP, WORLD RMU$OPEN Close RMU$OPEN WORLD Collect RMU$ANALYZE SYSPRV, BYPASS Optimizer_ Statistics Convert RMU$CONVERT, SYSPRV, BYPASS RMU$RESTORE Copy_Database RMU$COPY SYSPRV, BYPASS Delete RMU$ANALYZE SYSPRV, BYPASS Optimizer_ Statistics Dump After_ RMU$DUMP SYSPRV, BYPASS Journal Dump Areas RMU$DUMP SYSPRV, BYPASS Dump Backup_File RMU$DUMP, READ BYPASS RMU$BACKUP, RMU$RESTORE Dump Export READ BYPASS Dump Header RMU$DUMP, SYSPRV, BYPASS RMU$BACKUP, RMU$OPEN Dump Lareas RMU$DUMP SYSPRV, BYPASS Dump Recovery_ READ BYPASS Journal Dump Row Cache RMU$DUMP SYSPRV, BYPASS Dump Snapshots RMU$DUMP SYSPRV, BYPASS Dump Users RMU$DUMP, WORLD RMU$BACKUP, RMU$OPEN Extract RMU$UNLOAD SYSPRV, BYPASS Insert Optimizer RMU$ANALYZE SYSPRV, Statistics BYPASS Load RMU$LOAD SYSPRV, BYPASS Load Audit RMU$SECURITY SECURITY, BYPASS Load Plan RMU$LOAD SYSPRV, BYPASS Monitor Reopen_ WORLD, SETPRV Log CMKRNL, DETACH, PSWAPM, ALTPRI, SYSGBL, SYSNAM, SYSPRV, BYPASS Monitor Start WORLD, SETPRV CMKRNL, DETACH, PSWAPM, ALTPRI, PRMMBX, SYSGBL, SYSNAM, SYSPRV, BYPASS Monitor Stop WORLD, SETPRV CMKRNL, DETACH, PSWAPM, ALTPRI, PRMMBX, SYSGBL, SYSNAM, SYSPRV, BYPASS Move_Area RMU$MOVE SYSPRV, BYPASS Open RMU$OPEN WORLD Optimize After_ RMU$BACKUP, SYSPRV, BYPASS Journal RMU$RESTORE Reclaim RMU$ALTER SYSPRV, BYPASS Recover RMU$RESTORE SYSPRV, BYPASS Recover Resolve RMU$RESTORE SYSPRV, BYPASS Repair RMU$ALTER SYSPRV, BYPASS Resolve RMU$RESTORE SYSPRV, BYPASS Restore RMU$RESTORE SYSPRV, BYPASS Restore Only_ RMU$RESTORE SYSPRV, BYPASS Root Server After_ RMU$OPEN WORLD Journal Reopen_ Output Server After_ RMU$OPEN WORLD Journal Start Server After_ RMU$OPEN WORLD Journal Stop Server Backup_ RMU$OPEN WORLD Journal Resume Server Backup_ RMU$OPEN WORLD Journal Suspend Server Record_ RMU$OPEN WORLD Cache Set After_ RMU$ALTER, SYSPRV, BYPASS Journal RMU$BACKUP, RMU$RESTORE Set AIP RMU$DUMP SYSPRV, BYPASS Set Audit RMU$SECURITY SECURITY, BYPASS Set Buffer RMU$ALTER SYSPRV, BYPASS Object Set Corrupt_ RMU$ALTER, SYSPRV, BYPASS Pages RMU$BACKUP, RMU$RESTORE Set Galaxy RMU$ALTER SYSPRV, BYPASS Set Global RMU$ALTER SYSPRV, BYPASS Buffers Set Logminer RMU$ALTER, SYSPRV, BYPASS RMU$BACKUP, RMU$RESTORE Set Privilege RMU$SECURITY SECURITY, BYPASS Set Row_Cache RMU$ALTER SYSPRV, BYPASS Set Shared RMU$ALTER SYSPRV, BYPASS Memory Show After_ RMU$BACKUP, SYSPRV, BYPASS Journal RMU$RESTORE, RMU$VERIFY Show AIP RMU$DUMP SYSPRV, BYPASS Show Audit RMU$SECURITY SECURITY, BYPASS Show Corrupt_ RMU$BACKUP, SYSPRV, BYPASS Pages RMU$RESTORE, RMU$VERIFY Show Locks WORLD Show Optimizer_ RMU$ANALYZE, SYSPRV, BYPASS Statistics RMU$SHOW Show Privilege RMU$SECURITY SECURITY, BYPASS Show Statistics RMU$SHOW SYSPRV, BYPASS, WORLD Show System WORLD Show Users RMU$SHOW, WORLD RMU$BACKUP, RMU$OPEN Show Version Unload RMU$UNLOAD SYSPRV, BYPASS Unload After_ RMU$DUMP SYSPRV, BYPASS Journal Verify RMU$VERIFY SYSPRV, BYPASS 2 Alter Invokes the RdbALTER utility for Oracle Rdb. NOTE Oracle Corporation recommends that the RdbALTER utility be used only as a last resort to provide a temporary patch to a corrupt database. The RdbALTER utility should not be used as a routine database management tool. Use the RdbALTER utility only after you fully understand the internal data structure, know the information the database should contain, and know the full effects of the command. Because of the power of the RdbALTER utility and the cascading effects it can have, Oracle Corporation recommends that you experiment on a copy of the damaged database before applying the RdbALTER utility to a production database. To invoke the RdbALTER utility, enter the RMU Alter command in the following format: $ RMU/ALTER [root-file-spec] The optional root file parameter identifies the database you want to alter. If you specify this parameter, you automatically attach to the specified database. If you do not specify this parameter, you must use the RdbALTER ATTACH command. See the RdbALTER Help for more information on the ATTACH command. The RMU Alter command responds with the following prompt: RdbALTER> This prompt indicates that the system expects RdbALTER command input. To access the RdbALTER Help file, enter the following: RdbALTER> HELP To use the RMU Alter command for a database, you must have the RMU$ALTER privilege in the root file ACL for the database or the OpenVMS SYSPRV or BYPASS privilege. You must have the OpenVMS SYSPRV or BYPASS privilege if you are using an RMU Alter command to change a file name. 2 Analyze Displays information about stored and actual cardinality values for tables and indexes, database space utilization in the database, index structures for the database, or the accessibility through indexes of data records in the database. 3 Database Gathers and displays statistics on how the database uses storage, logical area, or page space. 4 Description The RMU Analyze command provides a maintenance tool for database administrators. It generates a formatted display of statistical information that describes storage utilization in the database. Information is displayed selectively for storage areas and logical areas, or for a range of pages in a storage area. You can use the RMU Analyze command to analyze the following: o Space utilization for database pages o Space utilization for storage areas o Space utilization for logical areas 4 Format (B)0RMU/Analyze root-file-spec Command Qualifiers x Defaults x /Areas[=storage-area-list] x /Areas /[No]Binary_Output=file-option-list x /Nobinary_Output /End=integer x /End=last-page /Exclude=(options) x No logical areas excluded /[No]Lareas[=logical-area-list] x /Lareas /Option = {Normal | Full | Debug} x /Option=Normal /Output=file-name x /Output=SYS$OUTPUT /Start = integer x /Start=first-page 4 Parameters 5 root-file-spec The file specification for the database root file to be analyzed. The default file extension is .rdb. 4 Command_Qualifiers 5 Areas Areas[=storage-area-list] Areas=* Specifies the storage areas to be analyzed. You can specify each storage area by name or by the area's ID number. The default, the Areas qualifier, results in analysis of all storage areas. You can also specify the Areas=* qualifier to analyze all storage areas. If you specify more than one storage area, separate the storage area names or ID numbers in the storage-area-list parameter with a comma and enclose the list in parentheses. If you omit the Areas qualifier, information for all the storage areas is displayed. You can use the Start and End qualifiers with the Areas qualifier to analyze specific pages. If you use the Start and End qualifiers when you specify more than one storage area in the storage-area-list parameter, the same specified range of pages are analyzed in each specified storage area. The Areas qualifier can be used with an indirect command file. See the Indirect-Command-Files help entry for more information. 5 Binary_Output Binary_Output=file-option-list Nobinary_Output Allows you to direct the summary results to a binary file, and to create a record definition file that is compatible with the data dictionary for the binary output file. The binary output file can be loaded into an Oracle Rdb database by using the RMU Load command with the Record_Definition qualifier for use by a user-written management application or procedure. The binary output can also be used directly by the user-written application or procedure. The valid file options are: o File=file-spec The File option causes the Analyze command data to be stored in an RMS file that contains a fixed-length binary record for each storage area and logical area analyzed. The default file extension for the binary output file is .unl. The following command creates the binary output file analyze_out.unl: $ RMU/ANALYZE/BINARY_OUTPUT=FILE=ANALYZE_OUT MF_PERSONNEL.RDB o Record_Definition=file-spec The Record_Definition option causes the Analyze command data record definition to be stored in an RMS file. The output file contains the definition in a subset of the data dictionary command format, a format very similar to RDO field and relation definitions. The default file extension for the record definition output file is .rrd. The following command creates the output file analyze_out.rrd: $ RMU/ANALYZE/BINARY_OUTPUT=RECORD_DEFINITION=ANALYZE_OUT - _$ MF_PERSONNEL.RDB You can specify both file options in one command by separating them with a comma and enclosing them within parentheses, for example: $ RMU/ANALYZE/BINARY_OUTPUT= - _$ (FILE=ANALYZE_OUT,RECORD_DEFINITION=ANALYZE_OUT) - _$ MF_PERSONNEL.RDB If you specify the Binary_Output qualifier, you must specify at least one of the options. The default is the Nobinary_Output qualifier, which does not create an output file. 5 End End=integer Specifies the ending page number for the analysis. The default is the end of the storage area file. 5 Exclude Exclude=System_Records Exclude=Metadata Exclude=(System_Records, Metadata) Excludes information from the RMU Analyze command output. You can specify Exclude=System_Records or Exclude=Metadata, or both. If you specify both options, enclose them within parentheses and separate each option with a comma. When you do not specify the Exclude qualifier, data is provided for all the logical areas in the database. The options are as follows: o System_Records Information on the RDB$SYSTEM_RECORDS logical areas is excluded from the Analyze command output. o Metadata Information on all the Oracle Rdb logical areas (for example, the RDB$SYSTEM_RECORDS and RDB$COLLATIONS_NDX logical areas) is excluded from the RMU Analyze command output. Data is accumulated for the logical areas excluded with the Exclude qualifier, but the data is excluded from the Analyze output. You cannot use the Exclude qualifier and the Lareas qualifier in the same RMU Analyze command. 5 Lareas Lareas[=logical-area-list] Lareas=* Nolareas Specifies the logical areas to be analyzed. Each table in the database is associated with a logical area name. The default, the Lareas qualifier, results in analysis of all logical areas. You can also specify the Lareas=* qualifier to analyze all logical areas. If you specify more than one logical area name, separate the logical area names in the logical-area-list with a comma and enclose the list in parentheses. The Lareas qualifier can be used with indirect command files. See the Indirect-Command-Files help entry for more information. 5 Option Option=Normal Option=Full Option=Debug Specifies the type of information and level of detail the analysis will include. Three types of output are available: o Normal Output includes only summary information. The Normal option is the default. o Full Output includes histograms and summary information. o Debug Output includes internal information about the data, as well as histograms and summary information. In general, use the Debug option for diagnostic support purposes. You can also use the Debug option to extract data and perform an independent analysis. 5 Output Output=file-name Specifies the name of the file where output will be sent. The default file extension is .lis. If you do not specify the Output qualifier, the output is sent to SYS$OUTPUT. 5 Start Start=integer Specifies the starting page number for the analysis. The default is 1. 4 Usage_Notes o To use the RMU Analyze command for a database, you must have the RMU$ANALYZE privilege in the root file ACL for the database or the OpenVMS SYSPRV or BYPASS privilege. o When the RMU Analyze command is issued for a closed database, the command executes without other users being able to attach to the database. o Detected asynchronous prefetch should be enabled to achieve the best performance of this command. Beginning with Oracle Rdb V7.0, by default, detected asynchronous prefetch is enabled. You can determine the setting for your database by issuing the RMU Dump command with the Header qualifier. If detected asynchronous prefetch is disabled, and you do not want to enable it for the database, you can enable it for your Oracle RMU operations by defining the following logicals at the process level: $ DEFINE RDM$BIND_DAPF_ENABLED 1 $ DEFINE RDM$BIND_DAPF_DEPTH_BUF_CNT P1 P1 is a value between 10 and 20 percent of the user buffer count. o The following RMU Analyze command directs the results into a record definition file called db.rrd. This file is compatible with the syntax for creating new columns and tables in the data dictionary. $ RMU/ANALYZE/BINARY_OUTPUT=RECORD_DEFINITION=DB.RRD MF_PERSONNEL $! Display the db.rrd file created by the previous command: $ TYPE DB.RRD DEFINE FIELD RMU$DATE DATATYPE IS DATE. DEFINE FIELD RMU$AREA_NAME DATATYPE IS TEXT SIZE IS 32. DEFINE FIELD RMU$STORAGE_AREA_ID DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$FLAGS DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$TOTAL_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$EXPANDED_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$FRAGMENTED_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$EXPANDED_FRAGMENT_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$TOTAL_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$FRAGMENTED_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$FRAGMENT_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$PAGE_LENGTH DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$MAX_PAGE_NUMBER DATATYPE IS SIGNED LONGWORD. DEFINE FIELD RMU$FREE_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$OVERHEAD_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$AIP_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$ABM_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$SPAM_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$INDEX_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$BTREE_NODE_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$HASH_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATES_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$OVERFLOW_BYTES DATATYPE IS F_FLOATING. DEFINE FIELD RMU$LOGICAL_AREA_ID DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$RELATION_ID DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$RECORD_ALLOCATION_SIZE DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$TOTAL_SPACE DATATYPE IS F_FLOATING. DEFINE RECORD RMU$ANALYZE_AREA. . . . o The following list describes each of the fields in the db.rrd record definition: - RMU$DATE Contains the date that the Analyze operation was done - RMU$AREA_NAME Contains the name of the storage area that was analyzed - RMU$STORAGE_AREA_ID Contains the area ID of the storage area that was analyzed - RMU$FLAGS The three possible values in this field have the following meanings: * 0-Indicates that the record is a storage area record, not a logical area record * 1-Indicates that data compression is not enabled for the logical area * 3-Indicates that data compression is enabled for the logical area - RMU$TOTAL_BYTES Contains the total size of the data stored in the logical area - RMU$EXPANDED_BYTES Contains the total size of the stored data in the logical area after decompression - RMU$FRAGMENTED_BYTES Contains the number of bytes in the stored fragments - RMU$EXPANDED_FRAGMENT_BYTES Contains the number of bytes in the stored fragments after decompression - RMU$TOTAL_COUNT Contains the total number of records stored - RMU$FRAGMENTED_COUNT - Contains the number of fragmented records - RMU$FRAGMENT_COUNT Contains the number of stored fragments - RMU$PAGE_LENGTH Contains the length in bytes of a database page in the storage area - RMU$MAX_PAGE_NUMBER Contains the page number of the last initialized page in the storage area - RMU$FREE_BYTES Contains the number of free bytes in the storage area - RMU$OVERHEAD_BYTES Contains the number of bytes used for overhead in the storage area - RMU$AIP_COUNT Contains the number of the area inventory pages (AIPs) in the storage area - RMU$ABM_COUNT Contains the number of area bit map (ABM) pages in the storage area - RMU$SPAM_COUNT Contains the number of space area management (SPAM) pages in the storage area - RMU$INDEX_COUNT Contains the number of index records in the storage area - RMU$BTREE_NODE_BYTES Contains the number of bytes for sorted indexes in the storage area - RMU$HASH_BYTES Contains the number of bytes for hashed indexes in the storage area - RMU$DUPLICATES_BYTES Contains the number of bytes for duplicate key values for sorted indexes in the storage area - RMU$OVERFLOW_BYTES Contains the number of bytes for hash bucket overflow records in the storage area - RMU$LOGICAL_AREA_ID Contains the logical area ID of the logical area that was analyzed - RMU$RELATION_ID Contains the record type of the row in the logical area that was analyzed - RMU$RECORD_ALLOCATION_SIZE Contains the size of a row when the table was initially defined - RMU$TOTAL_SPACE Contains the number of bytes available for storing user data in the logical area (used space + free space + overhead) 4 Examples Example 1 The following command analyzes the EMPIDS_LOW and EMP_INFO storage areas in the mf_personnel database: $ RMU/ANALYZE/AREAS=(EMPIDS_LOW,EMP_INFO)/OUTPUT=EMP.OUT - _$ MF_PERSONNEL.RDB Example 2 Both of the following commands analyze the DEPARTMENTS and SALARY_HISTORY storage areas in the mf_personnel database: $! Using storage area names to specify storage areas $ RMU/ANALYZE/AREAS=(DEPARTMENTS,SALARY_HISTORY) MF_PERSONNEL.RDB - $ /OUTPUT=DEP_SAL.OUT $! $! Using storage area ID numbers to specify storage areas $ RMU/ANALYZE/AREAS=(2,9) MF_PERSONNEL.RDB /OUTPUT=DEP_SAL.OUT 3 Cardinality Generates a formatted display of the actual and stored cardinality values for specified tables and indexes. Also, if the stored cardinality values are different from the actual cardinality values, the RMU Analyze Cardinality command allows you to update the stored cardinality values. NOTE Beginning in Oracle Rdb Version 7.0, the RMU Analyze Cardinality command has been deprecated and might be removed in future versions of Oracle Rdb. The features available through this command are now available through the RMU Collect Optimizer_Statistics command and the RMU Show Optimizer_Statistics command. In addition, updating cardinality information for indexes using the RMU Analyze Cardinality command may cause poor performance because the prefix cardinality information is not collected. Therefore, Oracle Corporation recommends that you use the RMU Collect Optimizer_Statistics and RMU Show Optimizer_ Statistics commands instead of the RMU Analyze Cardinality command. See Collect_Optimizer_Statistics and Show Optimizer_ Statistics for information on the RMU Collect Optimizer_ Statistics and the RMU Show Optimizer_Statistics commands. 4 Description The actual cardinality values for tables and indexes can be different from the stored cardinality values in your database's RDB$SYSTEM storage area if RDB$SYSTEM has been set to read- only access. When rows are added to or deleted from tables and indexes after the RDB$SYSTEM storage area has been set to read- only access, the cardinality values for these tables and indexes are not updated. For indexes, the cardinality value is the number of unique entries for an index that allows duplicates. If the index is unique, Oracle Rdb stores zero for the cardinality, and uses the table cardinality instead. For tables, the cardinality value is the number of rows in the table. Oracle Rdb uses the cardinality values of indexes and tables to influence decisions made by the optimizer. If the actual cardinality values of tables and indexes are different from the stored cardinality values, the optimizer's performance can be adversely affected. When you use the SQL ALTER DATABASE statement to set the RDB$SYSTEM storage area to read-only access for your database, the Oracle Rdb system tables in the RDB$SYSTEM storage area are also set to read-only access. When the Oracle Rdb system tables are set to read-only access: o Automatic updates to table and index cardinality are disabled. o Manual changes made to the cardinalities to influence the optimizer are not allowed. o The I/O associated with the cardinality update is eliminated. With the RMU Analyze Cardinality command, you can: o Display the stored and actual cardinality values for the specified tables and indexes. o Update the stored cardinality value for a specified table or index with either the actual value or an alternative value of your own choosing. Oracle Corporation recommends that you update the stored cardinality value with the actual cardinality value. Specifying a value other than the actual cardinality value can result in poor database performance. 4 Format (B)0RMU Analyze/Cardinality root-file-spec [table-or-index-name[,...]] Command Qualifiers x Defaults x /[No]Confirm x /Noconfirm /Output = file-name x /Output = SYS$OUTPUT /Transaction_Type=option x /Transaction_Type=Automatic /[No]Update x /Noupdate 4 Parameters 5 root-file-spec The name of the database root file for which you want information. The default file extension is .rdb. This parameter is required. 5 table-or-index-name table-or-index-name[,...] The name of the table or index for which you want information about cardinality. The default is all tables and all enabled indexes. If you want information about a disabled index, you must specify it by name. If you do not accept the default and instead specify a table name, the RMU Analyze Cardinality command and any qualifiers you specify will affect only the named table; the command will not result in a display or update (if the Update qualifier is specified) of the indexes associated with the table. This parameter is optional. An indirect file reference can be used. See the Indirect-Command-Files help entry for more information. 4 Command_Qualifiers 5 Confirm Confirm Noconfirm Specify the Confirm qualifier with the Update qualifier to gain more control over the update function. When you specify the Confirm qualifier, you are asked whether the update should be performed for each selected table or index whose stored cardinality value is different from its actual cardinality value. You can respond with YES, NO, QUIT, or an alternative value for the stored cardinality. Specifying YES means that you want to update the stored cardinality with the actual cardinality value. Specifying NO means that you do not want to update the stored cardinality value. Specifying QUIT aborts the RMU Analyze Cardinality command, rolls back any changes you made to stored cardinalities, and returns you to the operating system prompt. Specifying an alternative value updates the stored cardinality value with the alternative value. When you specify the Noconfirm qualifier, you are not given the option of updating stored cardinality values with an alternative value of your own choosing. Instead, the stored cardinality values that differ from the actual cardinality values are automatically updated with the actual cardinality values. The default is the Noconfirm qualifier. The Confirm and Noconfirm qualifiers are meaningless and are ignored if they are specified without the Update qualifier. 5 Output Output=file-name Specifies the name of the file where output will be sent. The default is SYS$OUTPUT. The default output file type is .lis, if you specify a file name. 5 Transaction_Type Transaction_Type=option Allows you to specify the transaction mode for the transactions used to perform the analyze operation. Valid options are: o Automatic o Read_Only o Noread_Only You must specify an option if you use this qualifier. If you do not specify any form of this qualifier, the Transaction_Type=Automatic qualifier is the default. This qualifier specifies that Oracle RMU is to determine the transaction mode used for the analyze operation. If any storage area in the database (including those not accessed for the analyze operation) has snapshots disabled, the transactions used for the analyze operation are set to read/write mode. Otherwise, the transactions are set to read-only mode. The Transaction_Type=Read_Only qualifier specifies the transactions used to perform the analyze operation be set to read-only mode. When you explicitly set the transaction type to read-only, snapshots need not be enabled for all storage areas in the database, but must be enabled for those storage areas that are analyzed. Otherwise, you receive an error and the analyze operation fails. You might select this option if not all storage areas have snapshots enabled and you are analyzing objects that are stored only in storage areas with snapshots enabled. In this case, using the Transaction_Type=Read_Only qualifier allows you to perform the analyze operation and impose minimal locking on other users of the database. The Transaction_Type=Noread_Only qualifier specifies that the transactions used to for the analyze operation be set to read/write mode. You might select this option if you want to eradicate the growth of snapshot files that occurs during a read- only transaction and are willing to incur the cost of increased locking that occurs during a read/write transaction. 5 Update Update Noupdate Specify the Update qualifier to update the stored cardinality values of tables and indexes. You can perform an update only when the stored cardinality values differ from the actual cardinality values. When updating cardinality values, Oracle Corporation recommends that you update the stored cardinality values with the actual cardinality values, not with an alternative value of your own choosing. Specifying a value other than the actual cardinality value can result in poor database performance. The default is the Noupdate qualifier. Using the Update qualifier allows you to update the stored cardinality values of the specified tables and indexes even when the RDB$SYSTEM storage area is designated for read-only access. If you have set the RDB$SYSTEM storage area to read-only access, Oracle RMU sets it to read/write during execution of the RMU Analyze Cardinality command with the Update qualifier. Oracle RMU resets the area to read-only when the operation completes. If you are updating the stored cardinality for a table or index, and a system failure occurs before the RDB$SYSTEM storage area is changed back to read-only access, use the SQL ALTER DATABASE statement to manually change the database back to read-only access. However, note that if you have set the area to read-only, the update operation specified with the Update qualifier commences only if the database is off line or the database is quiescent. If you specify a table name parameter with an RMU Analyze Cardinality command that includes the Update qualifier, the associated indexes are not updated; you must specify each table and index you want to be updated or accept the default (by not specifying any table or index names) and have all items updated. Oracle Corporation recommends that you use the Update qualifier during offline operations or during a period of low update activity. If you update a cardinality while it is changing (as a result of current database activity), the end result is unpredictable. Specify the Noupdate qualifier when you want to display the stored and actual cardinality values only for the specified tables and indexes. 4 Usage_Notes o To use the RMU Analyze Cardinality command for a database, you must have the RMU$ANALYZE privilege in the root file ACL for the database or the OpenVMS SYSPRV or BYPASS privilege. o You must have the SQL ALTER privilege for the database to update a read-only RDB$SYSTEM storage area. o If you specify a name for the table-or-index-name parameter that is both an index name and a table name, the RMU Analyze Cardinality command performs the requested operation for both the table and index. o Although you can alter the cardinality of a unique index using the RMU Analyze Cardinality command, it has no effect. (A unique index has only unique keys and does not have any duplicate keys.) Because the cardinality of a unique index and the table it indexes are the same, Oracle Rdb uses the table cardinality value when performing operations that involve the cardinality of a unique index. Oracle Rdb does not use the cardinality value stored for a unique index, nor does it attempt to update this value as rows are stored or deleted. o When the RMU Analyze Cardinality command is issued for a closed database, the command executes without other users being able to attach to the database. 4 Examples Example 1 The following command provides information on the cardinality for all indexes and tables in the sample mf_personnel database: $ RMU/ANALYZE/CARDINALITY/NOUPDATE MF_PERSONNEL.RDB /OUTPUT=CARD.LIS Example 2 The following command provides information on the cardinality for the EMPLOYEES table in the mf_personnel database: $ RMU/ANALYZE/CARDINALITY/NOUPDATE MF_PERSONNEL.RDB EMPLOYEES - _$ /OUTPUT=EMP.LIS 3 Indexes Generates a formatted display of statistical information that describes the index structures for the database. 4 Description The RMU Analyze Indexes command provides a maintenance tool for analyzing index structures and generates a formatted display of this statistical information. Information is displayed selectively for storage areas and logical areas, or for a range of pages in a storage area. You can use the RMU Analyze Indexes command to analyze the structures of both sorted (including ranked sorted) and hashed indexes. The following shows sample output from the RMU Analyze Index command: $ RMU/ANALYZE/INDEXES MF_PERSONNEL.RDB JH_EMPLOYEE_ID_RANKED ---------------------------------------------------------------------------- Indices for database - RDBVMS_DISK1:[DB]MF_PERSONNEL.RDB; ---------------------------------------------------------------------------- Index JH_EMPLOYEE_ID_RANKED for relation JOB_HISTORY duplicates allowed Max Level: 3, Nodes: 34, Used/Avail: 8693/13532 (64%), Keys: 133, Records: 0 Duplicate nodes:0, Used/Avail: 0/0 (0%), Keys: 100, Maps: 100, Records:4113 Total Comp/Uncomp IKEY Size: 600/798, Compression Ratio: .75 ---------------------------------------------------------------------------- Data included in the statistics display includes the following information: o The first line of output identifies the database in which the analyzed index resides. o The second line of output: - Specifies if the index is a hashed index. In the example, the index is not hashed, so the term hashed does not appear. - The index name - Whether or not duplicates are allowed. o Third line of output: - Max Level The maximum number of levels in the index. - Nodes The total number of nodes in the index. - Used/Avail (%) The number of bytes used by the index/the number of bytes available. (The percentage of space used by the index.) - Keys The sum of the dbkeys that point directly to data records plus those that point to duplicate nodes. - Records The number of data records to which the Keys (in the previous list item) point directly. o The fourth line of output: - Duplicate nodes For hashed and nonranked sorted indexes, this is the number of duplicate nodes in the index. For a ranked sorted index, this is the number of overflow nodes. With ranked sorted indexes, Oracle Rdb compresses duplicates using a byte- aligned bitmap compression. It compresses the list of dbkeys that point to duplicates and stores that list in the index key node. Oracle Rdb creates overflow nodes when the compressed list of duplicates does not fit in one index key node. This overflow node contains a bitmap compressed list of dbkeys and pointers to the next overflow node. Therefore, for ranked sorted indexes, the duplicate nodes count (overflow nodes) can be zero (0) if the compressed list of dbkeys that point to duplicates fits into one node. - Used/Avail (%) The number of bytes used by duplicate nodes/number of bytes available in the duplicate nodes. (The percentage of space used within the duplicate nodes of the index.) This value can be zero (0) for a ranked sorted index if the number of duplicate nodes is zero. - Keys The total number of dbkeys that point to a duplicate node or that point to the beginning of a duplicate node chain in the index. - Maps (appears only if the index is a ranked sorted index) The number of duplicate key data record bit maps used by ranked sorted indexes to represent the duplicate index key data record dbkeys. - Records The total number of data records pointed to by duplicate nodes. If the index is a ranked sorted index, Records refers to the number of data records pointed to by duplicate bit maps. o The fifth line of output (appears only if the index is compressed): - Total Comp/Uncomp IKEY Size The total byte count of the compressed leaf index keys (level 1 nodes only)/the total byte count that would be consumed if the index were not compressed - Compression ratio. The calculated ratio of Total Comp/Uncomp. A compression ratio greater than 1.0 indicates that the compressed index keys occupy more space than the uncompressed index keys. For more information on RMU Analyze Indexes and the display of index keys, refer to the Oracle Rdb7 Guide to Database Performance and Tuning. 4 Format (B)0RMU/Analyze/Indexes root-file-spec [index-name[,...]] Command Qualifiers x Defaults x /[No]Binary_Output[=file-option-list] x /Nobinary_Output /Exclude = Metadata x All index data displayed /Option = {Normal | Full | Debug} x /Option=Normal /Output = file-name x /Output=SYS$OUTPUT /Transaction_Type=option x /Transaction_Type=Automatic 4 Parameters 5 root-file-spec The file specification for the database root file for which you want information. The default file extension is .rdb. This parameter is required. 5 index-name index-name[,...] The name of the index for which you want information. The default is all enabled indexes. If you want information about a disabled index, you must specify it by name. This parameter is optional. An indirect file reference can be used. See the Indirect-Command- Files help entry for more information. The wildcard characters "%" and "*" can be used in the index name specification. The following examples demonstrate various combinations of use of the wildcard characters. $ RMU /ANALYZE /INDEX MF_PERSONNEL EMP* $ RMU /ANALYZE /INDEX MF_PERSONNEL *LAST%NAME $ RMU /ANALYZE /INDEX MF_PERSONNEL EMP%LAST%NAME $ RMU /ANALYZE /INDEX MF_PERSONNEL *HASH, *LAST* 4 Command_Qualifiers 5 Binary_Output Binary_output=file-option-list Nobinary_Output Specifying the Binary_Output qualifier allows you to store the summary results in a binary file, and to create a record definition file that is compatible with the data dictionary for the binary output file. The binary output can be loaded into an Oracle Rdb database by using the RMU Load command with the Record_Definition qualifier for use by a user-written management application or procedure. The binary output can also be used directly by the user-written application or procedure. The valid file options are: o File=file-spec The File option causes the RMU Analyze Indexes command data to be stored in an RMS file that contains a fixed-length binary record for each index analyzed. The default file extension for the binary output file is .unl. The following command creates the binary output file analyze_ out.unl: $ RMU/ANALYZE/INDEXES - _$ /BINARY_OUTPUT=FILE=ANALYZE_OUT MF_PERSONNEL.RDB o Record_Definition=file-spec The Record_Definition option causes the RMU Analyze Indexes command data record definition to be stored in an RMS file. The output file contains the record definition in a subset of the data dictionary command format. The default file extension for the record definition output file is .rrd. Refer to the rrd_file_syntax help topic for a description of the .rrd files. The following command creates the output file analyze_ out.rrd: $ RMU/ANALYZE/INDEXES - _$ /BINARY_OUTPUT=RECORD_DEFINITION=ANALYZE_OUT MF_PERSONNEL.RDB You can specify both file options in one command by separating them with a comma and enclosing them within parentheses, as follows: $ RMU/ANALYZE/INDEXES/BINARY_OUTPUT= - _$ (FILE=ANALYZE_OUT,RECORD_DEFINITION=ANALYZE_OUT) - _$ MF_PERSONNEL.RDB If you specify the Binary_Output qualifier, you must specify at least one of the options. The default is the Nobinary_Output qualifier, which does not create an output file. 5 Exclude Exclude=Metadata Excludes information from the RMU Analyze Indexes command output. When you specify the Exclude=Metadata qualifier, information on the Oracle Rdb indexes (for example, the RDB$NDX_REL_NAME_NDX and RDB$COLLATIONS_NDX indexes) is excluded from the RMU Analyze Indexes command output. When you do not specify the Exclude qualifier, data is provided for all indexes in the database. Data is accumulated for the indexes excluded with the Exclude qualifier, but the data is excluded from the RMU Analyze Indexes command output. You cannot specify the Exclude qualifier and one or more index names in the same RMU Analyze Indexes command. 5 Option Option=type Specifies the type of information and the level of detail the analysis will include. Three types of output are available: o Normal Output includes only summary information. The Normal option is the default. o Full Output includes histograms and summary information. This option displays a summary line for each sorted index level. o Debug Output includes internal information about the data, histograms, and summary information. Note the following when using this option to analyze compressed index keys: - The key lengths are from the compressed index keys. - The hexadecimal output for the keys is that of the uncompressed index keys. - The output includes summary statistics about the compressed index keys. In general, use the Debug option for diagnostic support purposes. You can also use the Debug option to extract data and perform an independent analysis. 5 Output Output=file-name Specifies the name of the file where output will be sent. The default is SYS$OUTPUT. The default output file extension is .lis, if you specify a file name. 5 Transaction_Type Transaction_Type=option Allows you to specify the transaction mode for the transactions used to perform the analyze operation. Valid options are: o Automatic o Read_Only o Noread_Only You must specify an option if you use this qualifier. If you do not use any form of this qualifier, the Transaction_ Type=Automatic qualifier is the default. This qualifier specifies that Oracle RMU is to determine the transaction mode used for the analyze operation. If any storage area in the database (including those not accessed for the analyze operation) has snapshots disabled, the transactions used for the analyze operation are set to read/write mode. Otherwise, the transactions are set to read-only mode. The Transaction_Type=Read_Only qualifier specifies the transactions used to perform the analyze operation be set to read-only mode. When you explicitly set the transaction type to read-only, snapshots need not be enabled for all storage areas in the database, but must be enabled for those storage areas that are analyzed. Otherwise, you receive an error and the analyze operation fails. You might select this option if not all storage areas have snapshots enabled and you are analyzing objects that are stored only in storage areas with snapshots enabled. In this case, using the Transaction_Type=Read_Only qualifier allows you to perform the analyze operation and impose minimal locking on other users of the database. The Transaction_Type=Noread_Only qualifier specifies that the transactions used to for the analyze operation be set to read/write mode. You might select this option if you want to eradicate the growth of snapshot files that occurs during a read- only transaction and are willing to incur the cost of increased locking that occurs during a read/write transaction. 4 Usage_Notes o To use the RMU Analyze Indexes command for a database, you must have the RMU$ANALYZE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o When the RMU Analyze Indexes command is issued for a closed database, the command executes without other users being able to attach to the database. o The following RMU Analyze Indexes command produces an RMS record definition file called index.rrd that can be read by the RMU Load command and the data dictionary: $ RMU/ANALYZE/INDEX/BINARY_OUTPUT=RECORD_DEFINITION=INDEX.RRD - _$ MF_PERSONNEL $! $! Display the index.rrd file created by the previous command: $ TYPE INDEX.RRD DEFINE FIELD RMU$DATE DATATYPE IS DATE. DEFINE FIELD RMU$INDEX_NAME DATATYPE IS TEXT SIZE IS 32. DEFINE FIELD RMU$RELATION_NAME DATATYPE IS TEXT SIZE IS 32. DEFINE FIELD RMU$LEVEL DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$FLAGS DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$USED DATATYPE IS F_FLOATING. DEFINE FIELD RMU$AVAILABLE DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATE_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATE_USED DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATE_AVAILABLE DATATYPE IS F_FLOATING. DEFINE FIELD RMU$KEY_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DATA_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATE_KEY_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATE_DATA_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$TOTAL_COMP_IKEY_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$TOTAL_IKEY_COUNT DATATYPE IS F_FLOATING. DEFINE RECORD RMU$ANALYZE_INDEX. o The following list describes each of the fields in the index.rrd record definition: - RMU$DATE Contains the date that the analyze operation was done - RMU$INDEX_NAME Contains the name of the index that was analyzed - RMU$RELATION_NAME Contains the name of the table for which the index is defined - RMU$LEVEL Contains the maximum number of index levels - RMU$FLAGS The eight possible values in this field have the following meanings: * 0-Index is sorted and not unique. A full report is not generated. * 1-Index is sorted and unique. A full report is not generated. * 2-Index is hashed and not unique. A full report is not generated. * 3-Index is hashed and unique. A full report is not generated. * 4-Index is sorted and not unique. A full report is generated. * 5- Index is sorted and unique. A full report is generated. * 6- Index is hashed and not unique. A full report is generated. * 7-Index is hashed and unique. A full report is generated. * 8-Index is sorted ranked and not unique. A full report is not generated. * 9-Index is sorted ranked and unique. A full report is not generated. * 12-Index is sorted ranked and not unique. A full report is generated. * 13-Index is sorted ranked and unique. A full report is generated. The RMU Analyze Indexes command uses the RMU$FLAGS bits shown in RMU$FLAGS Bits Used by the RMU Analyze Indexes Command for describing specific index information. Table 2 RMU$FLAGS Bits Used by the RMU Analyze Indexes Command Bit Offset Meaning 0 Unique index if true 1 Hashed index if true 2 Full report record if true 3 Ranked index if true When RMU$FLAGS has bit 2 set it means that a full report is generated. A full report has records for each level of the index. - RMU$COUNT Contains the number of index nodes - RMU$USED Contains the amount of available space that is used - RMU$AVAILABLE Contains the amount of space available in the index records initially - RMU$DUPLICATE_COUNT Contains the number of duplicate records - RMU$DUPLICATE_USED Contains the amount of available space used in the duplicate records - RMU$DUPLICATE_AVAILABLE Contains the amount of space available in the duplicate records initially - RMU$KEY_COUNT Contains the number of keys - RMU$DATA_COUNT Contains the number of records - RMU$DUPLICATE_KEY_COUNT Contains the number of duplicate keys - RMU$DUPLICATE_DATA_COUNT Contains the number of duplicate records - RMU$TOTAL_COMP_IKEY_COUNT Contains the number of compressed index key bytes - RMU$TOTAL_IKEY_COUNT Contains the number of bytes that would be used by index keys, had they not been compressed 4 Examples Example 1 The following command analyzes the JH_EMPLOYEE_ID and SH_ EMPLOYEE_ID indexes in the mf_personnel database: $ RMU/ANALYZE/INDEXES MF_PERSONNEL.RDB JH_EMPLOYEE_ID,SH_EMPLOYEE_ID - _$ /OUTPUT=EMP_ID_INDEX.LIS Example 2 The following commands demonstrate the differences you see when you analyze a nonranked sorted index and a ranked sorted index. Note the differences in the values for the Duplicate nodes. The nonranked sorted index displays 80 duplicate nodes. The ranked sorted index (before more duplicates are added) displays 0 duplicate nodes for the same data. After hundreds of more duplicates are added, the ranked sorted index shows only 3 duplicate nodes. The differences you see are because of the different way duplicate records are stored for nonranked sorted indexes and ranked sorted indexes. See the Description help entry under this command for details on these differences. $ ! Analyze a nonranked sorted index: $ ! $ RMU/ANALYZE/INDEXES MF_PERSONNEL.RDB JH_EMPLOYEE_ID ---------------------------------------------------------------------------- Indices for database - USER1:[DB]MF_PERSONNEL.RDB;1 ---------------------------------------------------------------------------- Index JH_EMPLOYEE_ID for relation JOB_HISTORY duplicates allowed Max Level: 2, Nodes: 4, Used/Avail: 768/1592 (48%), Keys: 103, Records: 20 Duplicate nodes: 80, Used/Avail: 2032/4696 (43%), Keys: 80, Records: 254 ---------------------------------------------------------------------------- $ ! Analyze a ranked sorted index defined on the same column as the $ ! nonranked sorted index: $ RMU/ANALYZE/INDEXES MF_PERSONNEL.RDB JH_EMPLOYEE_ID_RANKED ---------------------------------------------------------------------------- Indices for database - USER1:[DB]MF_PERSONNEL.RDB;1 ---------------------------------------------------------------------------- Index JH_EMPLOYEE_ID_RANKED for relation JOB_HISTORY duplicates allowed Max Level: 2, Nodes: 11, Used/Avail: 2318/4378 (53%), Keys: 110, Records: 20 Duplicate nodes: 0, Used/Avail: 0/0 (0%), Keys: 80, Maps: 80, Records: 254 ---------------------------------------------------------------------------- $ ! $ ! Insert many duplicates and analyze the ranked sorted index again: $ ! $ RMU/ANALYZE/INDEXES MF_PERSONNEL.RDB JH_EMPLOYEE_ID_RANKED ---------------------------------------------------------------------------- Indices for database - USER1:[DB]MF_PERSONNEL.RDB;1 ---------------------------------------------------------------------------- Index JH_EMPLOYEE_ID_RANKED for relation JOB_HISTORY duplicates allowed Max Level: 2, Nodes: 13, Used/Avail: 2705/5174 (52%), Keys: 112, Records: 20 Duplicate nodes:3, Used/Avail:850/1194 (71%), Keys:80, Maps: 83, Records:2964 ---------------------------------------------------------------------------- 3 Placement Generates a formatted display of statistical information describing the row placement relative to the index structures for the database. 4 Description The RMU Analyze Placement command provides a maintenance tool for analyzing row placement relative to index structures and generates a formatted display of this statistical information. Information is displayed selectively for any specified storage area. You can use the RMU Analyze Placement command to determine: o The maximum and average path length to a data record. (The maximum and average number of records touched to reach a data record.) o The estimated maximum I/O path length to a data record. o The estimated minimum I/O path length to a data record. o The frequency distributions for the database key (dbkey) path lengths, maximum I/O path lengths, and minimum I/O path lengths for specified indexes. o The distribution of data records on data pages in a storage area by logical area identifier (ID) and dbkey, the number of dbkeys needed to reach each data record, the maximum and minimum I/O path lengths needed to reach the data record, and the specific dbkey for the data record. 4 Format (B)0RMU/Analyze/Placement root-file-spec [index-name[,...]] Command Qualifiers x Defaults x /Areas[=storage-area-list] x /Areas /[No]Binary_Output[=file-option-list] x /Nobinary_Output /Exclude = Metadata x All index data displayed /Option = {Normal | Full | Debug} x /Option = Normal /Output=file-name x /Output = SYS$OUTPUT /Transaction_Type=option x /Transaction_Type=Automatic 4 Parameters 5 root-file-spec The file specification for the database root file to be analyzed. The default file extension is .rdb. 5 index-name index-name[,...] The name of the index for which you want information. The default is all enabled indexes. If you want information about a disabled index, you must specify it by name. This parameter is optional. An indirect file reference can be used. 4 Command_Qualifiers 5 Areas Areas[=storage-area-list] Areas=* Specifies the storage areas to be analyzed. You can specify each storage area by name or by the area's ID number. If you are interested in the placement information for a particular index, specify the area where the data resides, not where the index resides. For example, if you are interested in the placement information for the SH_EMPLOYEE_ID index of the mf_personnel database, you should specify SALARY_HISTORY as the storage area (which is where the data resides), not RDB$SYSTEM (which is where the index resides). If you do not specify the Areas qualifier, or if you specify the Areas qualifier but do not provide a storage-area-list, information for all the storage areas is displayed. If you specify more than one storage area, separate the storage area names or ID numbers in the storage-area-list with a comma and enclose the list within parentheses. If you specify more than one storage area with the Areas qualifier, the analysis Oracle RMU provides is a summary for all the specified areas. The analysis is not broken out into separate sections for each specified storage area. To get index information for a specific storage area, issue the RMU Analyze Placement command, specifying only that area with the Areas qualifier. The Areas qualifier can be used with an indirect file reference. See the Indirect-Command-Files help entry for more information. The Areas qualifier (without a storage-area-list) is the default. 5 Binary_Output Binary_Output[=file-option-list] Nobinary_Output Specifying the Binary_Output qualifier allows you to store the summary results in a binary file, and to create a record definition file that is compatible with the data dictionary for the binary output file. The binary output file can be loaded into an Oracle Rdb database by using the RMU Load command with the Record_Definition qualifier that can then be used by a user- written management application or procedure. The binary output can also be used directly by the user-written application or procedure. The valid file options are: o File=file-spec The File option causes the RMU Analyze Placement command data to be stored in an RMS file that contains a fixed-length binary record for each index analyzed. The default file extension for the binary output file is .unl. The following command creates the binary output file analyze_out.unl: $ RMU/ANALYZE/PLACEMENT - _$ /BINARY_OUTPUT=FILE=ANALYZE_OUT MF_PERSONNEL.RDB o Record_Definition=file-spec The Record_Definition option causes the RMU Analyze Placement command data record definition to be stored in an RMS file. The output file contains the record definition in a subset of the data dictionary command format. The default file extension for the record definition output file is .rrd. Refer to the rrd_file_syntax help topic for a description of .rrd files. The following command creates the output file analyze_out.rrd: $ RMU/ANALYZE/PLACEMENT - _$ /BINARY_OUTPUT=RECORD_DEFINITION=ANALYZE_OUT MF_PERSONNEL.RDB You can specify both file options in one command by separating them with a comma and enclosing them within parentheses, as follows: $ RMU/ANALYZE/PLACEMENT/BINARY_OUTPUT= - _$ (FILE=ANALYZE_OUT,RECORD_DEFINITION=ANALYZE_OUT) - _$ MF_PERSONNEL.RDB The default is the Nobinary_Output qualifier, which does not create an output file. 5 Exclude Exclude=Metadata Excludes information from the RMU Analyze Placement command data. When you specify the Exclude=Metadata qualifier, information on all the Oracle Rdb indexes (for example, the RDB$NDX_REL_NAME_NDX and RDB$COLLATIONS_NDX indexes) is excluded from the RMU Analyze Placement command output. When you do not specify the Exclude qualifier, data is provided for all indexes in the database. Data is accumulated for the indexes excluded with the Exclude qualifier, but the data is excluded from the RMU Analyze Placement command output. You cannot specify the Exclude qualifier and one or more index names in the same RMU Analyze Placement command. 5 Option Option=type Specifies the type of information and level of detail the analysis will include. Three types of output are available: o Normal Output includes only summary information. Normal is the default. o Full Output includes histograms and summary information. o Debug Output includes internal information about the data, histograms, and summary information. Output also displays uncompressed index keys from compressed indexes. The hexadecimal output is that of the uncompressed index key. However, the lengths shown are of the compressed index key. For more information on RMU Analyze Placement and the display of index keys, refer to the Oracle Rdb7 Guide to Database Performance and Tuning. 5 Output Output=file-name Specifies the name of the file where output will be sent. The default file type is .lis. If you do not specify the Output qualifier, the default output is SYS$OUTPUT. 5 Transaction_Type Transaction_Type=option Allows you to specify the transaction mode for the transactions used to perform the analyze operation. Valid options are: o Automatic o Read_Only o Noread_Only You must specify an option if you use this qualifier. If you do not use any form of this qualifier, the Transaction_ Type=Automatic qualifier is the default. This qualifier specifies that Oracle RMU is to determine the transaction mode used for the analyze operation. If any storage area in the database (including those not accessed for the analyze operation) has snapshots disabled, the transactions used for the analyze operation are set to read/write mode. Otherwise, the transactions are set to read-only mode. The Transaction_Type=Read_Only qualifier specifies the transactions used to perform the analyze operation be set to read-only mode. When you explicitly set the transaction type to read-only, snapshots need not be enabled for all storage areas in the database, but must be enabled for those storage areas that are analyzed. Otherwise, you receive an error and the analyze operation fails. You might select this option if not all storage areas have snapshots enabled and you are analyzing objects that are stored only in storage areas with snapshots enabled. In this case, using the Transaction_Type=Read_Only qualifier allows you to perform the analyze operation and impose minimal locking on other users of the database. The Transaction_Type=Noread_Only qualifier specifies that the transactions used for the analyze operation be set to read/write mode. You might select this option if you want to eradicate the growth of snapshot files that occurs during a read-only transaction and are willing to incur the cost of increased locking that occurs during a read/write transaction. 4 Usage_Notes o To use the RMU Analyze Placement command for a database, you must have the RMU$ANALYZE privilege in the root file ACL for the database or the OpenVMS SYSPRV or BYPASS privilege. o When the RMU Analyze Placement command is issued for a closed database, the command executes without other users being able to attach to the database. o The following RMU Analyze Placement command directs the results into an RMS record definition file called placement.rrd that is compatible with the data dictionary: $ RMU/ANALYZE/PLACEMENT/BINARY_OUTPUT=RECORD_DEFINITION=PLACEMENT.RRD - _$ MF_PERSONNEL $! $! Display the placement.rrd file created by the previous command: $ TYPE PLACEMENT.RRD DEFINE FIELD RMU$DATE DATATYPE IS DATE. DEFINE FIELD RMU$INDEX_NAME DATATYPE IS TEXT SIZE IS 32. DEFINE FIELD RMU$RELATION_NAME DATATYPE IS TEXT SIZE IS 32. DEFINE FIELD RMU$LEVEL DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$FLAGS DATATYPE IS SIGNED WORD. DEFINE FIELD RMU$COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATE_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$KEY_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATE_KEY_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DATA_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$DUPLICATE_DATA_COUNT DATATYPE IS F_FLOATING. DEFINE FIELD RMU$TOTAL_KEY_PATH DATATYPE IS F_FLOATING. DEFINE FIELD RMU$TOTAL_PAGE_PATH DATATYPE IS F_FLOATING. DEFINE FIELD RMU$TOTAL_BUFFER_PATH DATATYPE IS F_FLOATING. DEFINE FIELD RMU$MAX_KEY_PATH DATATYPE IS F_FLOATING. DEFINE FIELD RMU$MAX_PAGE_PATH DATATYPE IS F_FLOATING. DEFINE FIELD RMU$MIN_BUF_PATH DATATYPE IS F_FLOATING. DEFINE RECORD RMU$ANALYZE_PLACEMENT. o The following list describes each of the fields in the placement.rrd record definition: - RMU$DATE Contains the date that the analyze operation was done - RMU$INDEX_NAME Contains the name of the index that was analyzed - RMU$RELATION_NAME Contains the name of the table for which the index is defined - RMU$LEVEL Contains the maximum number of index levels - RMU$FLAGS The six possible values in this field have the following meanings: * 0-Index is a sorted and not unique index * 1-Index is sorted and unique * 2-Index is hashed and not unique * 3-Index is hashed and unique * 4-Index is ranked sorted and not unique * 5-Index is ranked sorted and unique The RMU Analyze Placement command uses the RMU$FLAGS bits shown in RMU$FLAGS Bits Used by the RMU Analyze Placement Command for describing specific index information. Table 3 RMU$FLAGS Bits Used by the RMU Analyze Placement Command Bit Offset Meaning 0 Unique index if true 1 Hashed index if true 2 Ranked sorted index if true - RMU$COUNT Contains the number of index nodes - RMU$DUPLICATE_COUNT Contains the number of duplicate records - RMU$KEY_COUNT Contains the number of keys - RMU$DUPLICATE_KEY_COUNT Contains the number of duplicate keys - RMU$DATA_COUNT Contains the number of records - RMU$DUPLICATE_DATA_COUNT Contains the number of duplicate records - RMU$TOTAL_KEY_PATH Contains the total number of keys touched to access all the records - RMU$TOTAL_PAGE_PATH Contains the total number of pages touched to access all the records - RMU$TOTAL_BUFFER_PATH Contains the total number of buffers touched to access all the records - RMU$MAX_KEY_PATH Contains the largest number of keys touched to access any of the records - RMU$MAX_PAGE_PATH Contains the largest number of pages touched to access any of the records - RMU$MIN_BUF_PATH Contains the smallest number of buffers touched to access any of the records 4 Examples Example 1 The following command provides information on row storage relative to the DEPARTMENTS_INDEX index of the sample personnel database: $ RMU/ANALYZE/PLACEMENT MF_PERSONNEL.RDB DEPARTMENTS_INDEX 2 Backup There are three RMU Backup commands, as follows: o An RMU Backup command without the After_Journal qualifier creates a database backup file. o An RMU Backup command with the After_Journal qualifier creates a backup of the after-image journal (.aij) file. The .aij can reside on disk or on tape. The RMU Backup command with the After_Journal qualifier supports a two-stage journaling technique that saves disk space and creates a backup journal on tape. o An RMU Backup command with the Plan qualifier allows you to execute a List_Plan previously created with a parallel backup operation. This form of the Backup command does not accept a database name as a parameter. Instead, it requires the name of a list plan. 3 Database Creates a backup copy of the database and places it in a file. If necessary, you can later use the RMU Restore command to restore the database to the condition it was in at the time of the backup operation. 4 Description The RMU Backup command copies information contained in a database to a file. It provides a number of options that allow you to determine the following: o Whether to perform a parallel backup operation. When you specify a parallel backup operation, you must back up to tape or multiple disks. The Parallel Backup Monitor allows you to monitor the progress of a parallel backup operation. o Whether to back up the database to disk or tape. o The extent (how much of the database) to back up. The backup operation uses a multithreaded process to optimize the performance of the backup operation. See the Oracle Rdb Guide to Database Maintenance for a complete description of how multithreading works. A parallel backup operation, in addition to using multithreaded processes, uses a coordinator executor and multiple worker executors (subprocesses) to enhance the speed of the backup operation. You can also direct each worker executor to run on a different node within a cluster to further enhance the speed of the operation. You must have Oracle SQL/Services installed and running to perform a parallel backup operation. See the Oracle Rdb Guide to Database Maintenance for information on when a parallel backup operation is most useful. Use the Parallel qualifier to indicate to Oracle RMU that you want to perform a parallel backup operation. Use the Noexecute and List_Plan qualifiers to generate a Backup plan file. A Backup plan file records the backup options and specifications you enter on the command line in a text file. You can edit this text file to fine-tune your parallel backup operation and execute it, as needed, with the RMU Backup Plan command. Use the Statistics option to the Parallel qualifier if you want to monitor the progress of the parallel backup operation with the Parallel Backup Monitor. See the description of the Parallel, List_Plan, and Noexecute qualifiers, and the RMU Backup Plan command for details. You cannot use the Parallel Backup Monitor to monitor the progress of a non-parallel backup operation. However, you can achieve a close approximation of this by specifying the Executor_ Count=1 and the Statistics options with the Parallel qualifier. This results in a parallel backup operation with one executor and one controller that you can monitor with the Parallel Backup Monitor. Both parallel and non-parallel backup operations allow you to perform different types of backup operations with respect to the portions of the database to be backed up, as described in RMU Backup Options. Table 4 RMU Backup Options Storage Area Selection Database Page Complete By-Area Selection (All Areas) (Selected Areas) Full Copies the database root Copies the database (.rdb) file and all the root (.rdb) file and database pages in all backs up only the the storage areas in the database pages in the database. This is the storage areas that you default backup operation. specify on the backup Note that you must use command line. All the this type of backup prior storage areas in the to upgrading to a newer database are backed version of Oracle Rdb. up only if you specify Because this is the them all (or perform default operation, no a full and complete qualifiers are needed to backup operation). Use specify a full backup. the Include or Exclude qualifiers to specify the storage areas for a full by-area backup operation. Incremental Copies all database pages Copies the database that have been updated root (.rdb) file and since the latest full only the database backup operation and pages for the the database root file. specified storage Use the Incremental (or areas that have Incremental=Complete) changed since the qualifier to specify an latest full backup incremental and complete operation. Use the backup operation. Include or Exclude qualifier along with the Incremental=By_ Area qualifier to specify an incremental, by-area, backup operation. Oracle Corporation recommends that you use a full backup operation to back up a database if you have made changes in the physical or logical design. Performing an incremental backup operation under these circumstances can lead to the inability to recover the database properly. If you choose to perform a by-area backup operation, your database can be fully recovered after a system failure only if after-image journaling is enabled on the database. If your database has both read/write and read-only storage areas but does not have after-image journaling enabled, you should do complete backup operations (backup operations on all the storage areas in the database) at all times. Doing complete backup operations when after-image journaling is not enabled ensures that you can recover the entire database to its condition at the time of the previous backup operation. When a full backup file is created for one or more storage areas, the date and time of the last full backup file created for those storage areas (as recorded in the backup (.rbf) file) is updated. You can display the date and time of the last full backup operation on each of the storage areas in a database by executing an RMU Dump command with the Header qualifier on the latest backup (.rbf) file for the database. The date and time displayed by this command is the date and time of the last full backup operation performed for the area. Note that an incremental backup operation on a storage area does not update the date and time for the last full backup operation performed on the storage area that is recorded in the backup file. In the event of subsequent damage to the database, you can specify backup files in an RMU Restore command to restore the database to the condition it was in when you backed it up. The RMU Backup command writes backup files in compressed format to save space. Available or free space in the database root (.rdb) file and on each database page in a storage area (.rda) file is not written to the backup file. NOTE Use only the RMU Backup command to back up all Oracle Rdb databases. Do not back up a database by using any other method (such as the DCL BACKUP command). The database root of a database is updated only when the RMU Backup command is used. For detailed information on backing up a database to tape, see the Oracle Rdb Guide to Database Maintenance. 4 Format (B)0RMU/Backup root-file-spec backup-file-spec Command Qualifiers x Defaults /[No]Accept_Label x /Noaccept_Label /[No]Acl x /Acl /Active_IO=max-writes x /Active_IO=3 /Allocation=blocks x None /Block_Size=integer x See description /[No]Checksum_Verification x /Checksum_Verification /[No]Compression[=options] x /Nocompression /Crc[=Autodin_II] x See description /Crc=Checksum x See description /Nocrc x See description /[No]Database_Verification x /Database_Verification /Density=(density-value,[No]Compaction) x See description /Disk_File[=options] x None /Encrypt=({Value=|Name=}[,Algorithm=]) x See description /Exclude[=storage-area[,...] ] x See description /[No]Execute x See description /Extend_Quantity=number-blocks x /Extend_Quantity=2048 /[No]Group_Size=interval x See description /Include[=storage-area[,...] ] x See description (B)0/[No]Incremental x /Noincremental /Incremental={By_area|Complete} x None /Journal=file-name x See description /Label=(label-name-list) x See description /Librarian[=options] x None /List_Plan=output-file x See description /Loader_Synchronization[=Fixed] x See description /Lock_Timeout=seconds x See description /[No]Log[=Brief|Full] x Current DCL verify switch /Master x See description /[No]Media_Loader x See description /No_Read_Only x See description /[No]Record x Record /[No]Online x /Noonline /Owner=user-id x See description /Page_Buffers=number-buffers x /Page_Buffers=2 /Parallel=(Executor_Count=n[,options]) x See description /Prompt={Automatic|Operator|Client} x See description /Protection[=file-protection] x See description /[No]Quiet_Point x /Quiet_Point (B)0/Reader_Thread_Ratio=integer x See description /Restore_Options=file-name x None /[No]Rewind x /Norewind /[No]Scan_Optimization x See description /Tape_Expiration=date-time x The current time /Threads=n x See description 4 Parameters 5 root-file-spec The name of the database root file. The root file name is also the name of the database. The default file extension is .rdb. 5 backup-file-spec The file specification for the backup file. The default file extension is .rbf. Depending on whether you are performing a backup operation to magnetic tape, disk, or multiple disks, the backup file specification should be specified as follows: o If you are backing up to magnetic tape - Oracle Corporation recommends that you supply a backup file name that is 17 or fewer characters in length. File names longer than 17 characters might be truncated. See the Usage_Notes help entry under this command for more information about backup file names that are longer than 17 characters. - If you use multiple tape drives, the backup-file-spec parameter must be provided with (and only with) the first tape drive name. Additional tape drive names must be separated from the first and subsequent tape drive names with commas. See the Oracle Rdb Guide to Database Maintenance for more information about using multiple tape drives. o If you are backing up to multiple or single disk files - It is good practice to write backup files to a device other than the devices where the database root, storage area, and snapshot files of the database are located. This way, if there is a problem with the database disks, you can still restore the database from a backup file. - If you use multiple disk files, the backup-file-spec parameter must be provided with (and only with) the first disk device name. Additional disk device names must be separated from the first and subsequent disk device names with commas. You must include the Disk_File qualifier. For example: $ RMU/BACKUP/DISK_FILE MF_PERSONNEL.RDB - _$ DEVICE1:[DIRECTORY1]MFP.RBF,DEVICE2:[DIRECTORY2] As an alternative to listing the disk device names on the command line (which, if you use several devices, can exceed the line-limit length for a command line), you can specify an options file in place of the backup-file-spec. For example: $ RMU/BACKUP/DISK_FILE LARGE_DB "@DEVICES.OPT" The contents of devices.opt might appear as follows: DEVICE1:[DIRECTORY1]LARGE_DB.RBF DEVICE2:[DIRECTORY2] The resulting backup files created from such an options file would be: DISK1:[DIRECTORY1]LARGE_DB.RBF DISK2:[DIRECTORY2]LARGE_DB01.RBF Note that the same directory must exist on each device before you issue the command. Also, if you forget to specify the Disk_File qualifier, you receive an error message similar to the following: $ RMU/BACKUP MF_PERSONNEL DEVICE1:[DIRECTORY1]MFP.RBF, - _$ DEVICE2:[DIRECTORY2] %RMU-F-NOTBACFIL, DEVICE1:[DIRECTORY1]MFP.RBF; is not a valid backup file %RMU-F-FTL_BCK,Fatal error for BACKUP operation at 2-MAY-2001 09:44:57.04 4 Command_Qualifiers 5 Accept_Label Accept_Label Specifies that RMU Backup should keep the current tape label it finds on a tape during a backup operation even if that label does not match the default label or that specified with the Label qualifier. Operator notification does not occur unless the tape's protection, owner, or expiration date prohibit writing to the tape. However, a message is logged (assuming logging is enabled) and written to the backup journal file (assuming you have specified the Journal qualifier) to indicate that a label is being preserved and which drive currently holds that tape. This qualifier is particularly useful when your backup operation employs numerous previously used (and thus labeled) tapes and you want to preserve the labels currently on the tapes. However, you are responsible for remembering the order in which tapes were written. For this reason, it is a good idea to use the Journal qualifier when you use the Accept_Label qualifier. If you do not specify this qualifier, the default behavior of RMU Backup is to notify the operator each time it finds a mismatch between the current label on the tape and the default label (or the label you specify with the Label qualifier). See the description of the Labels qualifier under this command for information on default labels. See How Tapes are Relabeled During a Backup Operation in the Usage_Notes help entry under this command for a summary of which labels are applied under a variety of circumstances. 5 Acl Acl Noacl Specifies whether to back up the root file access control list (ACL) for a database when you back up the database. The root file ACL controls users privileges to issue Oracle RMU commands. If you specify the Acl qualifier, the root file ACL will be backed up with the database. If you specify the Noacl qualifier, the root file ACL will not be backed up with the database. The Noacl qualifier can be useful if you plan to restore the database on a system where the identifiers in the current root file ACL will not be valid. The default is the Acl qualifier. 5 Active_IO Active_IO=max-writes Specifies the maximum number of write operations to a backup device that the RMU Backup command will attempt simultaneously. This is not the maximum number of write operations in progress; that value is the product of active system I/O operations and the number of devices being written to simultaneously. The value of the Active_IO qualifier can range from 1 to 5. The default value is 3. Values larger than 3 can improve performance with some tape drives. 5 Allocation Allocation=blocks Specifies the size, in blocks, which the backup file is initially allocated. The minimum value for the number-blocks parameter is 1; the maximum value allowed is 2147483647. If you do not specify the Allocation_Quantity qualifier, the Extend_Quantity value effectively controls the file's initial allocation. This qualifier cannot be used with backup operations to tape. 5 Block_Size Block_Size=integer Specifies the maximum record size for the backup file. The size can vary between 2048 and 65,024 bytes. The default value is device dependent. The appropriate block size is a compromise between tape capacity and error rate. The block size you specify must be larger than the largest page length in the database. 5 Checksum_Verification Checksum_Verification Nochecksum_Verification The Checksum_Verification qualifier requests that the RMU Backup command verify the checksum stored on each database page before the backup operation is applied, thereby providing end-to-end error detection on the database I/O. The default value is Checksum_Verification. Oracle Corporation recommends that you accept this default behavior for your applications. The default behavior prevents you from including corrupt database pages in backup files and optimized .aij files. Without the checksum verifications, corrupt data pages in these files are not detected when the files are restored. The corruptions on the restored page may not be detected until weeks or months after the backup file is created, or it is possible the corruption may not be detected at all. The Checksum_Verification qualifier uses additional CPU resources but provides an extra measure of confidence in the quality of the data that is backed up. Note that if you specify the Nochecksum qualifier, and undetected corruptions exist in your database, the corruptions are included in your backup file and are restored when you restore the backup file. Such a corruption might be difficult to recover from, especially if it is not detected until long after the restore operation is performed. 5 Compression Compression=LZSS Compression=Huffman Compression=ZLIB=level Nocompression Allows you to specify the compression method to use before writing data to the backup file. This reduces performance, but may be justified when the backup file is a disk file, or is being backed up over a busy network, or is being backed up to a tape drive that does not do its own compression. You probably do not want to specify the Compression qualifier when you are backing up a database to a tape drive that does its own compression; in some cases doing so can actually result in a larger file. If you specify the Compression qualifier without a value, the default is COMPRESSION=ZLIB=6. The level value (ZLIB=level) is an integer between 1 and 9 specifying the relative compression level with one being the least amount of compression and nine being the greatest amount of compression. Higher levels of the compression use increased CPU time while generally providing better compression. The default compression level of 6 is a balance between compression effectiveness and CPU consumption. OLDER ORACLE RDB 7.2 RELEASES AND COMPRESSED RBF FILES Prior releases of Oracle Rdb are unable to read RBF files compressed with the ZLIB algorithm. In order to read compressed backups with Oracle Rdb 7.2 Releases prior to V7.2.1, they must be made with /COMPRESSION=LZSS or /COMPRESSION=HUFFMAN explicitly specified (because the default compression algorithm has been changed from LZSS to ZLIB). Oracle Rdb Version 7.2.1 is able to read compressed backups using the LZSS or HUFFMAN algorithms made with prior releases. 5 Crc[=Autodin_II] CRC[=AUTODIN_II] Uses the AUTODIN-II polynomial for the 32-bit cyclic redundancy check (CRC) calculation and provides the most reliable end-to-end error detection. This is the default for NRZ/PE (800/1600 bits/inch) tape drives. If you enter only Crc as the qualifier, RMU Backup assumes you are specifying Crc=Autodin_II. 5 Crc=Checksum Crc=Checksum Uses one's complement addition, which is the same computation used to do a checksum of the database pages on disk. This is the default for GCR (6250 bits/inch) tape drives and for TA78, TA79, and TA81 tape drives. The Crc=Checksum qualifier allows detection of data errors. 5 Nocrc Nocrc Disables end-to-end error detection. This is the default for TA90 (IBM 3480 class) drives. NOTE The overall effect of the Crc=Autodin_II, Crc=Checksum, and Nocrc qualifier defaults is to make tape reliability equal to that of a disk. If you retain your tapes longer than 1 year, the Nocrc default might not be adequate. For tapes retained longer than 1 year, use the Crc=Checksum qualifier. If you retain your tapes longer than 3 years, you should always use the Crc=Autodin_II qualifier. Tapes retained longer than 5 years could be deteriorating and should be copied to fresh media. See the Oracle Rdb Guide to Database Maintenance for details on using the Crc qualifiers to avoid underrun errors. 5 Database_Verification Database_Verification Nodatabase_Verification The RMU /BACKUP command performs a limited database root file verification at the start of the backup operation. This verification is intended to help prevent backing up a database with various detectable corruptions or inconsistancies of the root file or associated database structures. However, in some limited cases, it can be desirable to avoid these checks. The qualifier /NODATABASE_VERIFICATION may be specified to avoid the database root file verification at the start of the backup. The default behavior is /DATABASE_VERIFICATION. Oracle strongly recommends accepting the default of /DATABASE_VERIFICATION. 5 Density Density=(density-value,[No]Compaction) Specifies the density at which the output volume is to be written. The default value is the format of the first volume (the first tape you mount). You do not need to specify this qualifier unless your tape drives support data compression or more than one recording density. The Density qualifier is applicable only to tape drives. RMU Backup returns an error message if this qualifier is used and the target device is not a tape drive. If you specify a density value, RMU Backup assumes that all tape drives can accept that value. If your systems are running OpenVMS versions prior to 7.2-1, specify the Density qualifier as follows: o For TA90E, TA91, and TA92 tape drives, specify the number in bits per inch as follows: - Density = 70000 to initialize and write tapes in the compacted format. - Density = 39872 or Density = 40000 for the noncompacted format. o For SCSI (Small Computer System Interface) tape drives, specify Density = 1 to initialize and write tapes by using the drive's hardware data compression scheme. o For other types of tape drives you can specify a supported density value between 800 and 160000 bits per inch. o For all tape drives, specify Density = 0 to initialize and write tapes at the drive's standard density. Do not use the Compaction or NoCompaction keyword for systems running OpenVMS versions prior to 7.2-1. On these systems, compression is determined by the density value and cannot be specified. Oracle RMU supports the OpenVMS tape density and compression values introduced in OpenVMS Version 7.2-1. The following table lists the added density values supported by Oracle RMU. DEFAULT 800 833 1600 6250 3480 3490E TK50 TK70 TK85 TK86 TK87 TK88 TK89 QIC 8200 8500 8900 DLT8000 SDLT SDLT320 SDLT600 DDS1 DDS2 DDS3 DDS4 AIT1 AIT2 AIT3 AIT4 LTO2 LTO3 COMPACTION NOCOMPACTION If the OpenVMS Version 7.2-1 density values and the previous density values are the same (for example, 800, 833, 1600, 6250), the specified value is interpreted as an OpenVMS Version 7.2-1 value if the tape device driver accepts them, and as a previous value if the tape device driver accepts previous values only. For the OpenVMS Version 7.2-1 values that accept tape compression you can use the following syntax: /DENSITY = (new_density_value,[No]Compaction) In order to use the Compaction or NoCompaction keyword, you must use one of the following density values that accepts compression: DEFAULT 3480 3490E 8200 8500 8900 TK87 TK88 TK89 DLT8000 SDLT SDLT320 AIT1 AIT2 AIT3 AIT4 DDS1 DDS2 DDS3 DDS4 SDLT600 LTO2 LTO3 Refer to the OpenVMS documentation for more information about density values. 5 Disk_File Disk_File[=(options)] Specifies that you want to perform a multithreaded backup operation to disk files, floppy disks, or other disks external to the PC. You can use the following keywords with the Disk_File qualifier: o Writer_Threads Specifies the number of threads that Oracle RMU should use when performing a multithreaded backup operation to disk files. You can specify no more than one writer thread per device specified on the command line (or in the command parameter options file). By default, one writer thread is used. This qualifier and all qualifiers that control tape operations (Accept_Label, Density, Label, Loader_Synchronization, Master, Media_Loader, Rewind, and Tape_Expiration) are mutually exclusive. 5 Encrypt Encrypt=({Value=|Name=}[,Algorithm=]) The Encrypt qualifier encrypts the save set file of a database backup. Specify a key value as a string or, the name of a predefined key. If no algorithm name is specified the default is DESCBC. For details on the Value, Name and Algorithm parameters see HELP ENCRYPT. This feature requires the OpenVMS Encrypt product to be installed and licensed on this system. 5 Exclude Exclude[=storage-area[,...]] Specifies the storage areas that you want to exclude from the backup file. If you specify neither the Exclude nor the Include qualifier with the RMU Backup command, or if you specify the Exclude qualifier but do not specify a list of storage area names, a full and complete backup operation is performed on the database. This is the default behavior. If you specify a list of storage area names with the Exclude qualifier, RMU Backup excludes those storage areas from the backup file and includes all of the other storage areas. If you specify more than one database storage area in the Exclude qualifier, place a comma between each storage area name and enclose the list of names within parentheses. Use the Exclude=* qualifier to indicate that you want only the database root file to be backed up. Note that a backup file created with the Exclude=* qualifier can be restored only with the RMU Restore Only_Root command. You can use an indirect command file as shown in the following example: $ RMU/BACKUP/EXCLUDE="@EXCLUDE_AREAS.OPT" - _$ MF_PERSONNEL.RDB PARTIAL_MF_PERS.RBF %RMU-I-NOTALLARE, Not all areas will be included in this backup file See the Indirect-Command-Files help entry for more information on indirect command files. If you use the Exclude qualifier with a list of storage area names, your backup file will be a by-area backup file because the Exclude qualifier causes database storage areas to be excluded from the backup file. The following example shows the informational message you receive if you do not back up all of the areas in the database: %RMU-I-NOTALLARE, Not all areas will be included in this backup file By using the RMU Backup and RMU Restore commands, you can back up and restore selected storage areas of your database. This Oracle RMU backup and restore by-area feature is designed to: o Speed recovery when corruption occurs in some (not all) of the storage areas of your database o Reduce the time needed to perform backup operations because some data (data in read-only storage areas, for example) does not need to be backed up with every backup operation performed on the database If you plan to use the RMU Backup and RMU Restore commands to back up and restore only selected storage areas for a database, you should perform full and complete backup operations on the database at regular intervals. If you plan to back up and restore only selected storage areas of a database, Oracle Corporation also strongly recommends that you enable after-image journaling for the database. This ensures that you can recover all of the storage areas in your database if a system failure occurs. If you do not have after-image journaling enabled and one or more of the areas restored with the RMU Restore command are not consistent with the unrestored storage areas, Oracle Rdb does not allow any transaction to use the storage areas that are not consistent in the restored database. In this situation, you can return to a working database by restoring the database, using the backup file from the last full and complete backup operation of the database storage areas. However, any changes made to the database since the last full and complete backup operation are not recoverable. If you do have after-image journaling enabled, use the RMU Recover command (or the Restore command with the Recover qualifier) to apply transactions from the .aij file to storage areas that are not consistent after the RMU Restore command completes; that is, storage areas that are not in the same state as the rest of the restored database. You cannot use these areas until you recover the database. When the RMU Recover command completes, your database will be consistent and usable. Using the Exclude or Include qualifier gives you greater flexibility for your backup operations, along with increased file management and recovery complexity. Users of large databases might find the greater flexibility of the backup operation to be worth the cost of increased file management and recovery complexity. You cannot specify the Exclude=area-list and Include=area-list qualifiers in the same RMU Backup command. 5 Execute Execute Noexecute Use the Execute and Noexecute qualifiers with the Parallel and List_Plan qualifiers to specify whether or not the backup plan file is to be executed. The following list describes the effects of using the Execute and Noexecute qualifier: o Execute Creates, verifies, and executes a backup list plan o Noexecute Creates and verifies, but does not execute a backup list plan. The verification determines such things as whether the storage areas listed in the plan file exist in the database. The Execute and Noexecute qualifiers are only valid when the Parallel and List_Plan qualifiers are also specified. If you specify the Execute or Noexecute qualifier without the List_Plan and Parallel qualifiers, RMU Backup generates and verifies a temporary backup list plan, but then deletes the backup list plan and returns a fatal error message. By default, the backup plan file is executed when you issue an RMU Backup command with the Parallel and List_Plan qualifiers. 5 Extend_Quantity Extend_Quantity=number-blocks Sets the size, in blocks, by which the backup file can be extended. The minimum value for the number-blocks parameter is 1; the maximum value is 65535. If you do not specify the Extend_ Quantity qualifier, the default number of blocks by which an on-disk backup file can be extended is 2048 blocks. This qualifier cannot be used with backup operations to tape. 5 Group_Size Group_Size=interval Nogroup_Size Specifies the frequency at which XOR recovery blocks are written to tape. The group size can vary from 0 to 100. Specifying a group size of zero or specifying the Nogroup_Size qualifier results in no XOR recovery blocks being written. The Group_Size qualifier is only applicable to tape, and its default value is 10. RMU Backup returns an error message if this qualifier is used and the target device is not a tape device. 5 Include Include[=storage-area[,...]] Specifies storage areas that you want to include in the backup file. If you specify neither the Include nor the Exclude qualifier with the RMU Backup command, a full and complete backup operation is performed on the database by default. You can specify the Include=* qualifier to indicate that you want all storage areas included in the backup file, but this is unnecessary because this is the default behavior. The default behavior is performed also when you specify the Include qualifier without specifying a list of storage area names. If you specify a list of storage area names with the Include qualifier, Oracle RMU includes those storage areas in the backup operation and excludes all of the other storage areas. If you specify more than one database storage area in the Include qualifier, place a comma between each storage area name and enclose the list of names within parentheses. You cannot specify the Exclude=area-list and Include=area-list qualifiers in the same RMU Backup command. If you use the Include qualifier, your backup operation will be a by-area backup operation because the areas not specified with the Include qualifier are excluded from the backup file. If you do not back up all of the areas in the database, you receive the following informational message: %RMU-I-NOTALLARE, Not all areas will be included in this backup file By using the RMU Backup and RMU Restore commands, you can back up and restore selected storage areas of your database. This Oracle RMU backup and restore by area feature is designed to: o Speed recovery when corruption occurs in some (not all) of the storage areas of your database o Reduce the time needed to perform backup operations because some data (data in read-only storage areas, for example) does not need to be backed up with every backup operation performed on the database See the description of the Exclude qualifier for information on the implications of using these commands to back up and restore selected areas of your database. The Include qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 5 Incremental Incremental[=By_Area or Complete] Noincremental Determines the extent of the backup operation to be performed. The four possible options are: o Noincremental If you do not specify any of the possible Incremental qualifier options, the default is the Noincremental qualifier. With the Noincremental qualifier, a full backup operation is performed on the database. o Incremental If you specify the Incremental qualifier, an incremental backup of all the storage areas that have changed since the last full and complete backup operation on the database is performed. o Incremental=By_Area If you specify the Incremental=By_Area qualifier, an incremental backup operation is performed. The Incremental=By_ Area qualifier backs up those database pages that have changed in each selected storage area since the last full backup operation was performed on the area. The last full backup operation performed on the area is the later of the following: - The last full and complete backup operation performed on the database - The last full by-area backup operation performed on the area With an incremental by-area backup operation, each storage area backed up might contain changes for a different time interval, which can make restoring multiple storage areas more complex. o Incremental=Complete If you specify the Incremental=Complete qualifier, an incremental backup operation on all of the storage areas that have changed since the last full and complete backup operation on the database is performed. Selecting the Incremental=Complete qualifier is the same as selecting the Incremental qualifier. Following a full database backup operation, each subsequent incremental backup operation replaces all previous incremental backup operations. The following two messages are meant to provide an aid for designing more effective backup strategies. They are printed as part of the per-area summary statistics, and they provide a guide to the incremental benefit of the incremental operation: o "Est. cost to backup relative to a full backup is x.yy" o "Est. cost to restore relative to a full restore is x.yy" These estimates are only approximate and reflect the disk input/output (I/O) cost for the backup or restore operations of that area. Tape I/O, CPU, and all other costs are ignored. The disk I/O costs take into account the number of I/O operations needed and the requirement for a disk head seek to perform the I/O. Each disk type has its own relative costs-transfer rate, latency, seek time-and the cost of a given sequence of I/Os is also affected by competition for the disk by other processes. Consequently, the estimates do not translate directly into "clock time." But they should nevertheless be useful for determining the point at which the incremental operation is becoming less productive. The relative costs can vary widely, and can be much higher than 1.00. The actual cost depends on the number and location of the pages backed up. An incremental restore operation must always follow a full restore operation, so the actual estimate of restoring the area is actually 1.00 higher than reported when that full restore operation is accounted for. The guideline that Oracle Corporation recommends is, "Perform full backup operations when the estimated cost of a restore operation approaches 2.00." 5 Journal Journal=file-name Allows you to specify a journal file to be used to improve tape performance during a restore operation. (This is not to be confused with an after-image journal file.) As the backup operation progresses, RMU Backup creates the journal file and writes to it a description of the backup operation containing identification of the tape drive names and the tape volumes and their contents. The default file extension is .jnl. The journal file must be written to disk; it cannot be written to tape along with the backup file. (Although you can copy the disk file to tape after it is written, if desired.) This journal file is used with the RMU Restore and the RMU Dump Backup commands to optimize their tape utilization. 5 Label Label=(label-name-list) Specifies the 1- to 6-character string with which the volumes of the backup file are to be labeled. The Label qualifier is applicable only to tape volumes. You must specify one or more label names when you use the Label qualifier. If you do not specify the Label (or Accept_Label) qualifier, RMU Backup labels the first tape used for a backup operation with the first 6 characters of the backup file name. Subsequent default labels are the first 4 characters of the backup file name appended with a sequential number. For example, if your backup file is my_backup.rbf, the default tape labels are my_bac, my_ b01, my_b02, and so on. When you reuse tapes, RMU Backup compares the label currently on the tape to the label or labels you specify with the Label qualifier. If there is a mismatch between the existing label and a label you specify, RMU Backup sends a message to the operator asking if the mismatch is acceptable (unless you also specify the Accept_Labels qualifier). If desired, you can explicitly specify the list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas and enclose the list of names within parentheses. If you are reusing tapes be certain that you load the tapes so that the label RMU Backup expects and the label on each tape will match, or be prepared for a high level of operator intervention. Alternatively, you can specify the Accept_Label qualifier. In this case, the labels you specify with the Label qualifier are ignored if they do not match the labels currently on the tapes and no operator intervention occurs. If you specify fewer labels than are needed, RMU Backup generates labels based on the format you have specified. For example, if you specify Label=TAPE01, RMU Backup labels subsequent tapes as TAPE02, TAPE03, and so on up to TAPE99. Thus, many volumes can be preloaded in the cartridge stacker of a tape drive. The order is not important because RMU Backup relabels the volumes. An unattended backup operation is more likely to be successful if all the tapes used do not have to be mounted in a specific order. Once the backup operation is complete, externally mark the tapes with the appropriate label so that the order can be maintained for the restore operation. Be particularly careful if you are allowing RMU Backup to implicitly label second and subsequent tapes and you are performing an unattended backup operation. Remove the tapes from the drives in the order in which they were written. Apply labels to the volumes following the logic of implicit labeling (for example, TAPE02, TAPE03, and so on). Oracle recommends you use the Journal qualifier when you employ implicit labeling in a multidrive, unattended backup operation. The journal file records the volume labels that were written to each tape drive. The order in which the labels were written is preserved in the journal. Use the RMU Dump Backup command to display a listing of the volumes written by each tape drive. You can use an indirect file reference with the Label qualifier. See the Indirect-command-files help entry for more information. See How Tapes are Relabeled During a Backup Operation in the Usage_Notes help entry under this command for a summary of which labels are applied under a variety of circumstances. 5 Librarian Librarian=options Use the Librarian qualifier to back up files to data archiving software applications that support the Oracle Media Management interface. The backup file name specified on the command line identifies the stream of data to be stored in the Librarian utility. If you supply a device specification or a version number it will be ignored. You can use the Librarian qualifier for parallel backup operations. The Librarian utility should be installed and available on all nodes on which the parallel backup operation executes. The Librarian qualifier accepts the following options: o Writer_Threads=n Use the Writer_Threads option to specify the number of backup data streams to write to the Librarian utility. The value of n can be from 1 to 99. The default is one writer thread. Each writer thread for a backup operation manages its own stream of data. Therefore, each thread uses a unique backup file name. The unique names are generated by incrementing the number added to the end of the backup file name. For example, if you specify the following Oracle RMU Backup command: $RMU/ BACKUP /LIBRARIAN=(WRITER_THREADS=3) /LOG DB FILENAM.RBF The following backup file data stream names are generated: FILENAME.RBF FILENAME.RBF02 FILENAME.RBF03 Because each data stream must contain at least one database storage area, and a single storage area must be completely contained in one data stream, if the number of writer threads specified is greater than the number of storage areas, it is set equal to the number of storage areas. o Trace_file=file-specification The Librarian utility writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian utility. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian utility. The pre-determined values are : - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical_name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian utility can use to specify catalogs or archives where Oracle Rdb backup files are stored, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of any Oracle RMU command that accesses the Librarian utility. The following OpenVMS logical names must be defined for use with a Librarian utility before you execute an Oracle RMU backup or restore operation. Do not use the Logical_Names option provided with the Librarian qualifier to define these logical names. o RMU$LIBRARIAN_PATH This logical name must be defined so that the shareable Librarian image can be loaded and called by Oracle RMU backup and restore operations. The translation must include the file type (for example, .exe), and must not include a version number. The shareable Librarian image must be an installed (known) image. See the Librarian utility documentation for the name and location of this image and how it should be installed. For a parallel RMU backup, define RMU$LIBRARIAN_ PATH as a system-wide logical name so that the multiple processes created by a parallel backup can all translate the logical. $ DEFINE /SYSTEM /EXECUTIVE_MODE - _$ RMU$LIBRARIAN_PATH librarian_shareable_image.exe o RMU$DEBUG_SBT This logical name is not required. If it is defined, Oracle RMU will display debug tracing information messages from modules that make calls to the Librarian shareable image. For a parallel RMU backup, the RMU$DEBUG_SBT logical should be defined as a system logical so that the multiple processes created by a parallel backup can all translate the logical. The following lines are from a backup plan file created by the RMU Backup/Parallel/Librarian command: Backup File = MF_PERSONNEL.RBF Style = Librarian Librarian_trace_level = # Librarian_logical_names = (- logical_name_1=equivalence_value_1, - logical_name_2=equivalence_value_2) Writer_threads = # The "Style = Librarian" entry specifies that the backup is going to a Librarian utility. The "Librarian_logical_names" entry is a list of logical names and their equivalence values. This is an optional parameter provided so that any logical names used by a particular Librarian utility can be defined as process logical names before the backup or restore operation begins. For example, some Librarian utilities provide support for logical names for specifying catalogs or debugging. You cannot use device specific qualifiers such as Rewind, Density, or Label with the Librarian qualifier because the Librarian utility handles the storage meda, not Oracle RMU. 5 List_Plan List_Plan=output-file Specifies that RMU Backup should generate a backup plan file for a parallel backup operation and write it to the specified output file. A backup plan file is a text file that contains qualifiers that can be specified on the RMU Backup command line. Qualifiers that you do not specify on the command line appear as comments in the backup list plan file. In addition, the backup plan file specifies the worker executor names along with the system node, storage areas, and tape drives assigned to each worker executor. You can use the generated backup plan file as a starting point for building a parallel backup operation to tape that is tuned for your particular configuration. The output file can be customized and then used with the RMU Backup Plan command. See Backup Plan for details. If you specify the Execute qualifier with the List_Plan qualifier, the backup plan file is generated, verified, and executed. If you specify the Noexecute qualifier with the List_ Plan qualifier, the backup plan file is generated and verified, but not executed. By default, the backup plan file is executed. The List_Plan qualifier is only valid when the Parallel qualifier is also specified. 5 Loader_Synchronization Loader_Synchronization[=Fixed] Allows you to preload tapes and preserve tape order to minimize the need for operator support. When you specify the Loader_ Synchronization qualifier and specify multiple tape drives, the backup operation writes to the first set of tape volumes concurrently then waits until each tape in the set is finished before assigning the next set of tape volumes. This ensures that the tape order can be preserved in the event that a restore operation from these tapes becomes necessary. One disadvantage with using the Loader_Synchronization qualifier with the Label qualifier is that because not all tape threads back up equal volumes of data, some threads may not need a subsequent tape to back up the assigned volume of data. In order to preserve the tape order, operator intervention may be needed to load the tapes in stages as backup threads become inactive. Use the keyword Fixed to force the assignment of tape labels to the drives regardless of how many tapes each drive actually uses. The Loader_Synchronization qualifier does result in reduced performance. For maximum performance, no drive should remain idle, and the next identified volume should be placed on the first drive that becomes idle. However, because the order in which the drives become idle depends on many uncontrollable factors and cannot be predetermined, without the Loader_ Synchronization qualifier, the drives cannot be preloaded with tapes. (If you do not want to relabel tapes, you might find that the Accept_Label qualifier is a good alternative to using the Loader_Synchronization qualifier. See the description of the Accept_Label qualifier for details.) Because the cost of using the Loader_Synchronization qualifier is dependent on the hardware configuration and the system load, the cost is unpredictable. A 5% to 20% additional elapsed time for the operation is typical. You must determine whether the benefit of a lower level of operator support compensates for the loss of performance. The Loader_Synchronization qualifier is most useful for large backup operations. See the Oracle Rdb Guide to Database Maintenance for more information on using the Loader_Synchronization qualifier, including information on when this qualifier might lead to unexpected results, and details on how this qualifier interacts with other RMU Backup command qualifiers. For very large backup operations requiring many tape volumes, managing the physical marking of tape volumes can be difficult. In such a case, you might consider using a library or archiving to automatically manage tape labeling for you. 5 Lock_Timeout Lock_Timeout=seconds Determines the maximum time the backup operation will wait for the quiet-point lock and any other locks needed during online backup operations. When you specify the Lock_Timeout=seconds qualifier, you must specify the number of seconds to wait for the quiet-point lock. If the time limit expires, an error is signaled and the backup operation fails. When the Lock_Timeout=seconds qualifier is not specified, the backup operation will wait indefinitely for the quiet-point lock and any other locks needed during an online backup operation. The Lock_Timeout=seconds qualifier is ignored for offline backup operations. 5 Log Log Log=Brief Log=Full Nolog Specifies whether the processing of the command is reported to SYS$OUTPUT. Specify the Log qualifier to request that the progress of the restore operation be written to SYS$OUTPUT, or the Nolog qualifier to suppress this report. If you specify the Log=Brief option, which is the default if you use the Log option without a qualifier, the log contains the start and completion time of each storage area. If you specify the Log=Full option, the log also contains thread assignment and storage area statistics messages. If you do not specify the Log or the Nolog qualifier, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 5 Master Master Controls the assignment of tape drives to output threads by allowing you to specify a tape drive as a master tape drive. This is a positional qualifier specified with a tape drive. When the Master qualifier is used, it must be used on the first tape drive specified. When the Master qualifier is specified, all additional tape drives become slaves for that tape drive until the end of the command line, or until the next Master qualifier, whichever comes first. If you specify the Master qualifier (without also specifying the Loader_Synchronization qualifier) on sets of tape drives, each master/slave set of tape drives will operate independently of other master/slave sets. If the Master qualifier is used on a tape drive that is not physically a master tape drive, the output performance of the backup operation will decrease. See the Oracle Rdb Guide to Database Maintenance for complete details on the behavior of the master qualifier. 5 Media_Loader Media_Loader Nomedia_Loader Use the Media_Loader qualifier to specify that the tape device receiving the backup file has a loader or stacker. Use the Nomedia_Loader qualifier to specify that the tape device does not have a loader or stacker. By default, if a tape device has a loader or stacker, RMU Backup should recognize this fact. However, occasionally RMU Backup does not recognize that a tape device has a loader or stacker. Therefore, when the first backup tape fills, RMU Backup issues a request to the operator for the next tape, instead of requesting the next tape from the loader or stacker. Similarly, sometimes RMU Backup behaves as though a tape device has a loader or stacker when actually it does not. If you find that RMU Backup is not recognizing that your tape device has a loader or stacker, specify the Media_Loader qualifier. If you find that RMU Backup expects a loader or stacker when it should not, specify the Nomedia_Loader qualifier. 5 No_Read_Only No_Read_Only Allows you to specify that you do not want any of the read-only storage areas in your database to be backed up when you back up the database. If you do not specify the No_Read_Only qualifier, any read-only storage area not specified with the Exclude qualifier will be included in the backup file. The No_Read_Only qualifier allows you to back up a database with many read-only storage areas without having to type a long list of read-only storage area names with the Exclude qualifier. If you specify the No_Read_Only qualifier, read-only storage areas are not backed up even if they are explicitly listed by the Include qualifier. There is no Read_Only qualifier. 5 Record Record Norecord The Record qualifier is set by default. Using the Norecord qualifier allows you to avoid the modification of the database with recent backup information. Hence the database appears as if it had not been backed up at this time. The main purpose of this qualifier is to allow a backup of a Hot Standby database without modifying the database files. The Norecord qualifier can be negated with the Record qualifier. 5 Online Online Noonline Specifying the Online qualifier permits users running active transactions at the time the command is entered to continue without interruption (unless the Noquiet_Point qualifier is also specified). Any subsequent transactions that start during the online backup operation are permitted as long as the transactions do not require exclusive access to the database, a table, or any index structure currently being backed up. To perform an online database backup operation, snapshots (either immediate or deferred) must be enabled. You can use the Online qualifier with the Incremental or Noincremental qualifiers. If you use the default, the Noonline qualifier, users cannot be attached to the database. If a user has invoked the database and the RMU Backup command is entered with the Noonline qualifier (or without the Online qualifier), an Oracle RMU error results. For example: %RMU-I-FILACCERR, error opening database root file DB_DISK:MF_PERSONNEL.RDB;1 -SYSTEM-W-ACCONFLICT, file access conflict The offline backup process (specified with the Noonline qualifier) has exclusive access to the database and does not require snapshot (.snp) files in order to work. The snapshot files can be disabled when the Noonline qualifier is used. Oracle Corporation recommends that you close the database (with the RMU Close command) when you perform the offline backup operation on a large database. If the database was opened with the SQL OPEN IS MANUAL statement, the RMU Backup command will fail unless the RMU Close command is used. If the database was opened with the SQL OPEN IS AUTOMATIC statement, the RMU Backup command might fail if the activity level is high (that is, users might access the database before the database is taken off line). Issuing the RMU Close command can force the users out of the database and give the RMU Backup command a chance to start; however, although recommended, issuing the RMU Close command is not required in this case. Synonymous with the Owner qualifier. See the description of the Owner qualifier. 5 Owner Owner=user-id Specifies the owner of the tape volume set. The owner is the user who will be permitted to restore the database. The user-id parameter must be one of the following types of identifier: o A user identification code (UIC) in [group-name,member-name] alphanumeric format o A user identification code (UIC) in [group-number,member- number] numeric format o A general identifier, such as SECRETARIES o A system-defined identifier, such as DIALUP The Owner qualifier cannot be used with a backup operation to disk. When used with tapes, the Owner qualifier applies to all continuation volumes. The Owner qualifier applies to the first volume only if the Rewind qualifier is also specified. If the Rewind qualifier is not specified, the backup operation appends the file to a previously labeled tape, so the first volume can have a protection different from the continuation volumes. See the Oracle Rdb Guide to Database Maintenance for information on tape label processing. 5 Page_Buffers Page_Buffers=number-buffers Specifies the number of disk buffers assigned to each storage area thread. The range is 2 to 5 with a default of 2. The higher values speed up scans for changed pages during an incremental backup operation, but they exact a cost in memory usage and larger working set requirements. 5 Parallel Parallel=(Executor_Count=n[,options]) Specifies that you want to perform a parallel backup operation. When you issue an RMU Backup command with the parallel qualifier, RMU Backup generates a plan file. This plan file describes how the parallel backup operation should be executed. If you specify the Noexecute qualifier, the plan file is generated, but not executed. If you specify the Execute qualifier (or accept the default), the plan file is executed immediately after RMU Backup creates it. The Executor_Count specifies the number of worker executors you want to use for the parallel backup operation. The number of worker executors must be equal to or less than the number of tape drives you intend to use. If you specify Executor_Count=1, the result is a non-parallel backup operation that is executed using the parallel backup procedure, including creation of the plan file and a dbserver process. You can specify one, both, or none of the following options: o Node=(node-list) The Node=(node-list) option specifies the names of the nodes in the cluster where the worker executors are to run. If more than one node is specified, all nodes must be in the same cluster and the database must be accessible from all nodes in the cluster. In addition, for a backup operation across nodes in a cluster to be successful, whoever starts SQL/Services must have proxy access among all nodes involved in the backup operation (assuming you are using DECnet). For example, if you specify the Nodes=(NODE1, NODE2, NODE3) as an option to the Parallel qualifier, whomever started SQL/Services must have access from NODE1 to NODE2, NODE1 to NODE3, NODE2 to NODE1, NODE2 to NODE3, NODE3 to NODE1, and NODE3 to NODE2. Separate node names in the node-list with commas. If you do not specify the Nodes option, all worker executors run on the node from which the parallel backup plan file is executed. o Server_Transport=(DECnet|TCP) To execute a parallel backup operation, SQL/Services must be installed on your system. By default, the RMU Backup command uses DECnet to access SQL/Services; if DECnet is not available, RMU Backup tries to use TCP/IP. Use the Server_Transport option to set the default behavior such that RMU Backup tries TCP/IP first. You can also use the SQL_NETWORK_TRANSPORT_TYPE configuration parameter to modify the default behavior. See the Oracle Rdb Installation and Configuration Guide for details on setting the SQL_NETWORK_ TRANSPORT_TYPE configuration parameter. o Statistics Specifies that you want RMU Backup to gather statistics on the parallel backup operation for use with the Parallel Backup Monitor. You must invoke the Parallel Backup Monitor, a Windowing interface, to view these statistics. To execute a parallel backup operation, SQL/Services must be installed on your system. By default, the RMU Backup command uses DECnet to access SQL/Services; if DECnet is not available, RMU Backup tries to use TCP/IP. You can use the SQL_NETWORK_ TRANSPORT_TYPE configuration parameter to set the default behavior such that RMU Backup tries TCP/IP first. See the Oracle Rdb Installation and Configuration Guide for details on setting the SQL_NETWORK_TRANSPORT_TYPE configuration parameter. Note that during a parallel backup operation, all tape requests are sent to the Operator; the parallel backup operation does not send tape requests to the user who issues the Backup command. Therefore, you should issue the DCL REPLY/ENABLE=TAPES command from the terminal that serves the operator before issuing the RMU Backup command. 5 Prompt Prompt=Automatic Prompt=Operator Prompt=Client Specifies where server prompts are to be sent. When you specify Prompt=Automatic, prompts are sent to the standard input device, and when you specify Prompt=Operator, prompts are sent to the server console. When you specify Prompt=Client, prompts are sent to the client system. 5 Protection Protection[=file-protection] Specifies the system file protection for the backup file produced by the RMU Backup command. The default file protection varies, depending on whether you backup the file to disk or tape. This is because tapes do not allow delete or execute access and the SYSTEM account always has both read and write access to tapes. In addition, a more restrictive class accumulates the access rights of the less restrictive classes. If you do not specify the Protection qualifier, the default protection is as follows: o S:RWED,O:RE,G,W if the backup is to disk o S:RW,O:R,G,W if the backup is to tape If you specify the Protection qualifier explicitly, the differences in protection applied for backups to tape or disk as noted in the preceding paragraph are applied. Thus, if you specify Protection=(S,O,G:W,W:R), that protection on tape becomes (S:RW,O:RW,G:RW,W:R). 5 Quiet_Point Quiet_Point Noquiet_Point Allows you to specify that the database backup operation is to occur either immediately or when a quiet point for database activity occurs. A quiet point is defined as a point where no active update transactions are in progress in the database. Therefore, this qualifier is used with the Online qualifier. When you specify the Noquiet_Point qualifier, RMU Backup proceeds with the backup operation as soon as the RMU Backup command is issued, regardless of any update transaction activity in progress in the database. Because RMU Backup must acquire concurrent- read locks on all physical and logical areas, the backup operation will fail if there are any active transactions with exclusive locks on a storage area. However, once RMU Backup has successfully acquired all concurrent-read storage area locks it should not encounter any further lock conflicts. If a transaction that causes Oracle Rdb to request exclusive locks is started while the backup operation is proceeding, that transaction will either wait or receive a lock conflict error, but the RMU Backup command will continue unaffected. See the Usage_Notes help entry under this command for recommendations on using the Quiet_Point and Noquiet_Point qualifiers. The default is the Quiet_Point qualifier. 5 Reader_Thread_Ratio Reader_Thread_Ratio=integer This qualifier has been deprecated. Use the /Threads qualifier instead. 5 Restore_Options Restore_Options=file-name Generates an options file designed to be used with the Options qualifier of the RMU Restore command. If you specify a full backup operation, all the storage areas will be represented in the options file. If you specify a by-area backup operation, only those areas included in the backup will be represented in the options file. The Restore_Options file is created at the end of the backup operation. By default, a Restore_Options file is not created. If you specify the Restore_Options qualifier and a file, but not a file extension, RMU Backup uses an extension of .opt by default. 5 Rewind Rewind Norewind Specifies that the magnetic tape that contains the backup file will be rewound before processing begins. The tape will be initialized according to the Label and Density qualifiers. The Norewind qualifier is the default and causes the backup file to be created starting at the current logical end-of-tape (EOT). The Rewind and Norewind qualifiers are applicable only to tape devices. RMU Backup returns an error message if these qualifiers are used and the target device is not a tape device. 5 Scan_Optimization Scan_Optimization Noscan_Optimization Specifies whether or not RMU Backup should employ scan optimizations during incremental backup operations. By default, RMU Backup optimizes incremental backup operations by scanning regions of the database that have been updated since the last full backup operation. The identity of these regions is stored in the database. Only these regions need to be scanned for updates during an incremental backup operation. This provides a substantial performance improvement when database activity is sufficiently low. However, there is a cost in recording this information in the database. In some circumstances the cost might be too high, particularly if you do not intend to use incremental backup operations. The Scan_Optimization qualifier has different effects, depending on the type of backup operation you perform. In brief, you can enable or disable the scan optimization setting only when you issue a full offline backup command, and you can specify whether to use the data produced by a scan optimization only when you issue an incremental backup command. The following list describes this behavior in more detail: o During an offline full backup operation, you can enable or disable the scan optimization setting. Specify the Scan_Optimization qualifier to enable recording of the identities of areas that change after this backup operation completes. Specify the Noscan_Optimization qualifier to disable recording of the identities of areas that change after this backup operation completes. By default, the recording state remains unchanged (from the state it was in prior to execution of the Backup command) during a full backup operation. Note that specifying the Scan_Optimization or Noscan_ Optimization qualifier with an offline full backup operation has no effect on the backup operation itself, it merely allows you to change the recording state for scan optimization. o During an online full backup operation, the qualifier is ignored. The recording state for scan optimization remains unchanged (from the state it was in prior to execution of the Backup command). If you execute an online full backup operation and specify the Scan_Optimization or Noscan_Optimization qualifier, RMU Backup returns an informational message to indicate that the qualifier is being ignored. o During an incremental backup operation, the qualifier directs whether the scan optimization data (if recorded previously) will be used during the operation. If you specify the Scan_Optimization qualifier, RMU Backup uses the optimization if Oracle Rdb has been recording the regions updated since the last full backup operation. If you specify the Noscan_Optimization qualifier, RMU Backup does not use the optimization, regardless of whether Oracle Rdb has been recording the identity of the regions updated since the last full backup operation. You cannot enable or disable the setting for scan optimizations during an incremental backup operation. By default, the Scan_Optimization qualifier is used during incremental backup operations. 5 Tape_Expiration Tape_Expiration=date-time Specifies the expiration date of the backup (.rbf) file. Note that when RMU Backup reads a tape, it looks at the expiration date in the file header of the first file on the tape and assumes the date it finds in that file header is the expiration date for the entire tape. Therefore, if you are backing up an .rbf file to tape, specifying the Tape_Expiration qualifier only has meaning if the .rbf is the first file on the tape. You can guarantee that the .rbf file will be the first file on the tape by specifying the Rewind qualifier and overwriting any existing files on the tape. When the first file on the tape contains an expiration date in the file header, you cannot overwrite the tape before the expiration date unless you have the OpenVMS SYSPRV or BYPASS privilege. Similarly, when you attempt to restore a .rbf file from tape, you cannot perform the restore operation after the expiration date recorded in the first file on the tape unless you have the OpenVMS SYSPRV or BYPASS privilege By default, no expiration date is written to the .rbf file header. In this case, if the .rbf file is the first file on the tape, the tape can be overwritten immediately. If the .rbf file is not the first file on the tape, the ability to overwrite the tape is determined by the expiration date in the file header of the first file on the tape. You cannot explicitly set a tape expiration date for an entire volume. The volume expiration date is always determined by the expiration date of the first file on the tape. The Tape_Expiration qualifier cannot be used with a backup file written to disk. See the Oracle Rdb Guide to Database Maintenance for information on tape label processing. 5 Threads=number Threads=number Specifies the number of reader threads to be used by the backup process. RMU creates so called internal 'threads' of execution to read data from one specific storage area. Threads run quasi-parallel within the process executing the RMU image. Each thread generates its own I/O load and consumes resources like virtual address space and process quotas (e.g. FILLM, BYTLM). The more threads, the more I/Os can be generated at one point in time and the more resources are needed to accomplish the same task. Performance increases with more threads due to parallel activities which keeps disk drives busier. However, at a certain number of threads, performance suffers because the disk I/O subsystem is saturated and I/O queues build up for the disk drives. Also the extra CPU time for additional thread scheduling overhead reduces the overall performance. Typically 2-5 threads per input disk drive are sufficient to drive the disk I/O susbsystem at its optimum. However, some controllers may be able to handle the I/O load of more threads, for example disk controllers with RAID sets and extra cache memory. In a backup operation, one writer thread is created per output stream. An output stream can be either a tape drive, a disk file or, a media library manager stream. In addition, RMU creates a number of reader threads and their number can be specified. RMU assigns a subset of reader threads to writer threads. RMU calculates the assignment so that roughly the same amount of data is assigned to each output stream. By default, five reader threads are created for each writer thread. If the user has specified the number of threads, then this number is used to create the reader thread pool. RMU always limits the number of reader threads to the number of storage areas. A threads number of 0 causes RMU to create one thread per storage area which start to run all in parallel immediately. Even though this may sound like a good idea to improve performance, this approach suffers performance for databases with a larger number (>10) of storage areas. For a very large number of storage areas (>800), this fails due to hard limitations in system resources like virtual address space. For a backup operation, the smallest threads number you can specify is the number of output streams. This guarantees that each writer thread has at least one reader thread assigned to it and does not produce an empty save set. Using a threads number equal to the number of output streams generates the smallest system load in terms of working set usage and disk I/O load. Disk I/O subsystems most likely can handle higher I/O loads. Using a slightly larger value than the number of output streams (for example, assigning more reader threads to a writer thread) typically results in faster execution time. 4 Usage_Notes o To use the RMU Backup command for a database, you must have the RMU$BACKUP privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o If you attempt to back up an area with detected corruptions (or which has corrupt pages logged to the CPT), the backup operation fails immediately. If you attempt to back up an area that contains an undetected corruptions (a corruption that has not been logged to the CPT), the backup operation proceeds until a corruption is found. These undetected corruptions are found only if you specify the Checksum qualifier with the Backup command. o The following list provides usage information for parallel backup operations: - When performing a parallel backup operation, do not allocate or mount any tapes manually; this is done automatically by RMU Backup. - You can monitor the progress of a backup operation to tape on your Windows system using the Parallel Backup Monitor. - You can use the Parallel Backup Monitor to monitor the progress of a parallel backup operation to tape. Specify your backup operation using the Parallel qualifier with the Executor_Count=1 option to approximate a non- parallel backup operation. Non-parallel backup operations (backup commands without the Parallel qualifier) cannot be monitored with the Parallel Backup Monitor. - If a parallel backup operation is issued from a server node, then RMU Backup communicates with SQL/Services to start the Coordinator. SQL/Services creates a Coordinator process. - If a parallel backup operation is issued from a client node (for example, using RMUwin), then the same SQL/Services process that is created to execute client/server RMU Backup commands is used as the Coordinator process. - You cannot use the Storage Library System (SLS) for OpenVMS with an RMU parallel backup. o Logical area threshold information for storage areas with uniform page format is recorded in the backup file. See the Oracle Rdb SQL Reference Manual for more information on logical area threshold information. o See the Oracle Rdb Guide to Database Maintenance for information on determining the working set requirements for a non-parallel backup operation. o The following list provides usage information for the Quiet_ Point and Noquiet_Point qualifiers - If the operation stalls when you attempt a quiet-point Oracle RMU backup operation, it may be because another user is holding the quiet-point lock. In some cases, there is no way to avoid this stall. In other cases you may find the stall is caused by a user who has previously issued and completed a read/write transaction, and is currently running a read-only transaction. When this user started the read/write transaction, the process acquired the quiet- point lock. Ordinarily, such a process retains this lock until it detaches from the database. You can set the RDM$BIND_SNAP_QUIET_POINT logical name to control whether or not such a process retains the quiet- point lock. Set the value of the logical name to "1" so that all transactions hold the quiet point lock until a backup process requests it. Read-only transactions will not obtain the quiet point lock; only read/write transactions will obtain the quiet point lock. Set the value of the logical name to "0" so that read-only transactions always release the quiet point lock at the beginning of the transaction, regardless of the existence of a backup process. All modified buffers in the buffer pool have to be written to disk before the transaction proceeds. Applications that utilize the fast commit feature and often switch between read-only and read/write transactions within a single attach may experience performance degradation if the logical is defined to "0". Oracle recommends that you do not define the RDB$BIND_SNAP_ QUIET_POINT logical for most applications. - If you intend to use the Noquiet_Point qualifier with a backup procedure that previously specified the Quiet_ Point qualifier (or did not specify either the Quiet_ Point or Noquiet_Point qualifier), you should examine any applications that execute concurrently with the backup operation. You might need to modify your applications or your backup procedure to handle the lock conflicts that might occur when you specify Noquiet_Point. When you specify the Quiet_Point qualifier, the backup operation begins when a quiet point is reached. Other update transactions that are started after the database backup operation begins are prevented from executing until after the root file for the database has been backed up (the backup operation on the database storage areas begins after the root file is backed up). - When devising your backup strategy for both the database and the after-image journal files, keep in mind the trade- offs between performing quiet-point backup operations and noquiet-point backup operations. A noquiet-point backup operation is quicker than a quiet-point backup operation, but usually results in a longer recovery operation. Because transactions can span .aij files when you perform noquiet- point .aij backup operations, you might have to apply numerous .aij files to recover the database. In a worst- case scenario, this could extend back to your last quiet- point .aij or database backup operation. If you rarely perform quiet-point backup operations, recovery time could be excessive. One method you can use to balance these trade-offs is to perform regularly scheduled quiet-point .aij backup operations followed by noquiet-point database backup operations. (You could do the converse, but a quiet- point backup of the .aij file improves the performance of the recovery operation should such an operation become necessary.) Periodically performing a quiet-point .aij backup operation helps to ensure that your recovery time will not be excessive. o Do not add new logical areas in the context of an exclusive transaction during an online backup operation. When new logical areas are added during an online backup operation such that new records are physically placed in a location that the backup operation has not processed yet, Oracle Rdb returns the following error: %RMU-F-CANTREADDBS, error reading pages !UL:!UL-!UL Logical areas that cause this problem are created when you do either of the following: - Create a new table, start a transaction that reserves the new table in exclusive mode, and load the table with rows. - Create a new table, start a transaction that reserves the new table in exclusive mode, and create an index for the table. Creating new tables and populating them, or creating new indexes do not pose a problem if the table is not reserved in exclusive mode. o If you back up a database without its root file ACL (using the Noacl qualifier of the RMU Backup command, for example), a user who wants to restore the database must have the OpenVMS SYSPRV or BYPASS privilege. o You might receive the RMU-I-WAITOFF informational message when you try to back up your database if the database was manually opened with the RMU Open command and has not been manually closed with the RMU Close command. You also receive this message when you issue an RMU Close command with the Nowait qualifier and users are still attached to the database. To back up your database, you must have exclusive access to the database root file. This error message usually indicates that you do not have exclusive access to the database root file because the operating system still has access to it. If your database was manually opened with the RMU Open command, you should be able to gain exclusive access to the database root file by manually closing the database with an RMU Close command. You can also receive this error message when you attempt other operations for which you must have exclusive access to the database root file. The solution in those cases is to attempt the operation again, later. Until you have exclusive access to the database root file, meaning that no other user gained access to the database between the time you issued the command and the time the command takes effect, you cannot complete those operations. o Backup files are typically smaller in size than the actual database. They exclude free space and redundant structural information that can be reconstructed with a restore operation. However, backup files also contain some overhead to support the backup format. Compression factors range from approximately 1.2 to 3 depending on the organization and fullness of the database. The compression factor achieved for a given database is generally quite stable and usually only changes with structural or logical reorganization. Do not use the size of the backup file as an indication of the size of the database files. Use the RMU Analyze command to determine the actual data content. o Backup performance is strongly affected by the job priority of the process running it. For best performance, a backup operation should execute at interactive priority, even when it is operating as a batch job. o The following list contains information of interest if you are performing a backup operation to tape: - If you back up the database to tape, and you do not specify the Parallel qualifier, you must mount the backup media by using the DCL MOUNT command before you issue the RMU Backup command. The tape must be mounted as a FOREIGN volume. Like the OpenVMS Backup utility (BACKUP), the RMU Backup command performs its own tape label processing. This does not prohibit backing up an Oracle Rdb database to an RMS file on a Files-11 disk. When you specify the Parallel qualifier, you need not mount the backup media because the parallel executors allocate and mount the drive and labels for you. - When RMU Backup creates a multivolume backup file, you can only append data to the end of the last volume. You cannot append data to the end of the first or any intermediate volumes. - The RMU Backup command uses asynchronous I/O. Tape support provided includes support for multifile volumes, multivolume files, and multithreaded concurrent tape processing. - If you allow RMU Backup to implicitly label tapes and you are using a tape drive that has a display (for example, a TA91 tape drive), the label displayed is the original label on the tape, not the label generated by RMU Backup. - Oracle Corporation recommends that you supply a name for the backup file that is 17 or fewer characters in length. File names longer than 17 characters can be truncated. The system supports four file-header labels: HDR1, HDR2, HDR3, and HDR4. In HDR1 labels, the file identifier field contains the first 17 characters of the file name you supply. The remainder of the file name is written into the HDR4 label, provided that this label is allowed. If no HDR4 label is supported, a file name longer than 17 characters will be truncated. The following Oracle RMU commands are valid. The terminating period for the backup file name is not counted as a character, and the default file type of .rbf is assumed. Therefore, the system interprets the file name as wednesdays_backup, which is 17 characters in length: $ RMU/BACKUP/REWIND/LABEL=TAPE MF_PERSONNEL MUA0:WEDNESDAYS_BACKUP. $ RMU/RESTORE/REWIND/LABEL=TAPE MUA0:WEDNESDAYS_BACKUP. The following Oracle RMU commands create a backup file that cannot be restored. Because no terminating period is supplied, the system supplies a period and a file type of .rbf, and interprets the backup file name as wednesdays_ backup.rbf, which is 20 characters in length. RMU truncates the backup file name to wednesdays_backup. When you attempt to restore the backed up file, RMU assumes the default extension of .rbf and returns an error when it cannot find the file wednesdays_backup.rbf on tape. $ RMU/BACKUP/REWIND/LABEL=TAPE MF_PERSONNEL MUA0:WEDNESDAYS_BACKUP $ RMU/RESTORE/REWIND/LABEL=TAPE MUA0:WEDNESDAYS_BACKUP - See the Oracle Rdb Guide to Database Maintenance for information on the steps RMU Backup follows in tape label checking for the RMU Backup command. - The RMU Backup command works correctly with unlabeled or nonstandard formatted tapes when the Rewind qualifier is specified. However, tapes that have never been used or initialized, and nonstandard tapes sometimes produce errors that make OpenVMS mount attempts fail repeatedly. In this situation, RMU Backup cannot continue until you use the DCL INITIALIZE command to correct the error. - How Tapes are Relabeled During a Backup Operation summarizes the tape labeling behavior of RMU Backup under a variety of circumstances. For example, the last row of the table describes what labels are applied when you specify both the Label=back qualifier and the Accept_Label qualifier and all the tapes (except the second) are already labeled and used in the following order: aaaa, no label, bbbb, dddd, cccc. The table shows that these tapes will be relabeled in the following order, with no operator notification occurring: aaaa, back02, bbbb, dddd, eeee. How Tapes are Relabeled During a Backup Operation assumes the backup file name is mf_personnel.rbf: Table 5 How Tapes are Relabeled During a Backup Operation Qualifiers Current Resulting Specified Labels Labels Operator Notification Neither None Label mf_ mf_ nor per per Accept_ mf_ mf_ Label p05 p05 mf_ mf_ p06 p06 mf_ mf_ p02 p02 mf_ mf_ p03 p03 Neither All tapes except second tape Label aaaa mf_ nor no per Accept_ label mf_ Label bbbb p02 dddd mf_ cccc p03 mf_ p04 mf_ p05 Label=back All tapes except second aaaa back no back02 label back03 bbbb back04 dddd back05 cccc Label=(back01, All tapes except second back02) aaaa back01 no back02 label back03 bbbb back04 dddd back05 cccc Accept_ None Label aaaa aaaa no mf_ label p02 bbbb bbbb dddd dddd cccc cccc Accept_ Label, None Label=back aaaa aaaa no back02 label bbbb bbbb dddd dddd cccc cccc o When you use more than one tape drive for a backup operation, ensure that all of the tape drives are the same type (for example, all of the tape drives must be TA90s or TZ87s or TK50s). Using different tape drive types (for example, one TK50 and one TA90) for a single database backup operation may make database restoration difficult or impossible. Oracle RMU attempts to prevent you from using different tape drive densities during a backup operation but is not able to detect all invalid cases and expects that all tape drives for a backup are of the same type. As long as all of the tapes used during a backup operation can be read by the same type of tape drive during a restore operation, the backup is likely to be valid. This may be the case, for example, when you use a TA90 and a TA90E. Oracle Corporation recommends that, on a regular basis, you test your backup and recovery procedures and environment using a test system. You should restore the database and then recover using after-image journals (AIJs) to simulate failure recovery of the production system. Consult the Oracle Rdb Guide to Database Maintenance and the Oracle Rdb Guide to Database Design and Definition for additional information about Oracle Rdb backup and restore operations. o You should use the density values added in OpenVMS Version 7.2-1 for OpenVMS tape device drivers that accept them because previously supported values may not work as expected. If previously supported values are specified for drivers that support the OpenVMS Version 7.2-1 density values, the older values are translated to the Version 7.2-1 density values if possible. If the value cannot be translated, a warning message is generated, and the specified value is used. If you use density values added in OpenVMS Version 7.2-1 for tape device drivers that do not support them, the values are translated to acceptable values if possible. If the value cannot be translated, a warning message is generated and the density value is translated to the existing default internal density value (MT$K_DEFAULT). One of the following density-related errors is generated if there is a mismatch between the specified density value and the values that the tape device driver accepts: %RMU-E-DENSITY, TAPE_DEVICE:[000000]DATABASE.BCK; does not support specified density %RMU-E-POSITERR, error positioning TAPE_DEVICE: %RMU-E-BADDENSITY, The specified tape density is invalid for this device o If you want to use an unsupported density value, use the VMS INITIALIZE and MOUNT commands to set the tape density. Do not use the Density qualifier. o The density syntax used on the command can also be used in the plan file for the Parallel RMU backup to tape process. o Oracle Rdb cannot continue a single .rda file across multiple disks. This means that, during a multidisk backup operation, each device must have enough free space to hold the largest storage area in the database. If the storage areas are on stripe sets and are larger than any actual single disk, then the devices specified for the backup file must be striped also. It is not possible to indicate which storage area should be backed up to a given device. o Because data stream names representing the database are generated based on the backup file name specified for the Oracle RMU backup command, you must either use a different backup file name to store the next backup of the database to the Librarian utility or first delete the existing data streams generated from the backup file name before the same backup file name can be reused. To delete the existing data streams stored in the Librarian utility, you can use a Librarian management utility or the Oracle RMU Librarian/Remove command. o If you are backing up to multiple disk devices using thread pools, the following algorithm to assign threads is used by the backup operation: - The size of each area is calculated as the product of the page length in bytes multiplied by the highest page number used (maximum page number) for that area. - The area sizes are sorted by descending size and ascending device name. For internal processing reasons, the system area is placed as the first area in the first thread. - Each of the remaining areas is added to the thread that has the lowest byte count. The same algorithm is used for tape devices, but the areas are partitioned among writer threads, not disk devices. o The partitioning for backup to multiple disk devices is done by disk device, not by output thread, because there will typically be more disk devices than output threads, and an area cannot span a device. 4 Examples Example 1 The following command performs a full backup operation on the mf_ personnel database and displays a log of the session: $ RMU/BACKUP MF_PERSONNEL - _$ DISK2[USER1]MF_PERS_FULL_BU.RBF /LOG Example 2 To perform an incremental backup operation, include the Incremental qualifier. Assume a full backup operation was done late Monday night. The following command performs an incremental backup operation on the database updates only for the following day: $ RMU/BACKUP/INCREMENTAL MF_PERSONNEL.RDB - _$ $222$DUA20:[BCK]TUESDAY_PERS_BKUP/LOG Example 3 To back up the database while there are active users, specify the Online qualifier: $ RMU/BACKUP/ONLINE MF_PERSONNEL.RDB - _$ $222$DUA20:[BACKUPS]PERS_BU.RBF /LOG Example 4 The following RMU Backup command includes only the EMPIDS_ LOW and EMPIDS_MID storage areas in the backup file of the mf_personnel database. All the other storage areas in the mf_ personnel database are excluded from the backup file: $ RMU/BACKUP/INCLUDE=(EMPIDS_LOW,EMPIDS_MID) - _$ MF_PERSONNEL.RDB $222$DUA20:[BACKUPS]MF_PERS_BU.RBF Example 5 The following command backs up the mf_personnel database but not the root file ACL for the database: $ RMU/BACKUP/NOACL MF_PERSONNEL MF_PERS_NOACL Example 6 The following command backs up the mf_personnel database without waiting for a quiet point in the database: $ RMU/BACKUP/NOQUIET_POINT MF_PERSONNEL MF_PERS_NQP Example 7 The following command creates a journal file, pers_journal.jnl, and a backup file, pers_backup.rbf. $ RMU/BACKUP/JOURNAL=PERS_JOURNAL MF_PERSONNEL PERS_BACKUP Example 8 The following example backs up all the storage areas in the mf_ personnel database except for the read-only storage areas. $ RMU/BACKUP/NO_READ_ONLY MF_PERSONNEL MF_PERSONNEL_BU Example 9 The following example assumes that you are using multiple tape drives to do a large backup operation. By specifying the Loader_ Synchronization qualifier, this command does not require you to load tapes as each becomes full. Instead, you can load tapes on a loader or stacker and RMU Backup will wait until all concurrent tape operations have concluded for one set of tape volumes before assigning the next set of tape volumes. Using this example, you: 1. Verify the database. 2. Allocate each tape drive. 3. Manually place tapes BACK01 and BACK05 on the $111$MUA0: drive. 4. Manually place tapes BACK02 and BACK06 on the $222$MUA1: drive. 5. Manually place tapes BACK03 and BACK07 on the $333$MUA2: drive. 6. Manually place tapes BACK04 and BACK08 on the $444$MUA3: drive. 7. Mount the first volume. 8. Perform the backup operation. 9. Dismount the last tape mounted. (This example assumes it is on the $444$MUA3: drive.) 10. Deallocate each tape drive. $ RMU/VERIFY DB_DISK:[DATABASE]MF_PERSONNEL.RDB $ $ ALLOCATE $111$MUA0: $ ALLOCATE $222$MUA1: $ ALLOCATE $333$MUA2: $ ALLOCATE $444$MUA3: $ $ MOUNT/FOREIGN $111$MUA0: $ $ RMU/BACKUP /LOG/REWIND/LOADER_SYNCHRONIZATION - _$ /LABEL=(BACK01, BACK02, BACK03, BACK04, BACK05, - _$ BACK06, BACK07, BACK08) - _$ DB_DISK:[MFPERS]MF_PERSONNEL.RDB - _$ $111$MUA0:PERS_FULL_MAR30.RBF/Master, $222$MUA1: - _$ $333$MUA1:/MASTER, $444$MUA3 $ $ DISMOUNT $444$MUA3: $ $ DEALLOCATE $111$MUA0: $ DEALLOCATE $222$MUA1: $ DEALLOCATE $333$MUA2: $ DEALLOCATE $444$MUA4: Example 10 The following example generates a parallel backup plan file, but does not execute it. The result is a backup plan file. See the next example for a description of the plan file. $ RMU/BACKUP/PARALLEL=(EXEC=4, NODE=(NODE1, NODE2)) - _$ /LIST_PLAN=(PARTIAL.PLAN)/NOEXECUTE/INCLUDE=(RDB$SYSTEM, EMPIDS_LOW, - _$ EMPIDS_MID, EMPIDS_OVER, SALARY_HISTORY, EMP_INFO) - _$ /LABEL=(001, 002, 003, 004, 005, 006, 007, 008, 009) - _$ /CHECKSUM_VERIFICATION - _$ MF_PERSONNEL TAPE1:MF_PARTIAL.RBF, TAPE2:, TAPE3:, TAPE4: Example 11 The following display shows the contents of the plan file, PARTIAL.PLAN created in the preceding example. The following callouts are keyed to this display: 1 The Plan Parameters include all the parameters specified on the RMU BACKUP command line and all possible command qualifiers. 2 Command qualifiers that are not specified on the command line are represented as comments in the plan file. This allows you to edit and adjust the plan file for future use. 3 Command qualifiers that are explicitly specified on the command line are represented in the plan file as specified. 4 Executor parameters are listed for each executor involved in the backup operation. ! Plan created on 28-JUN-1996 by RMU/BACKUP. Plan Name = PARTIAL Plan Type = BACKUP Plan Parameters: 1 Database Root File = DISK1:[DB]MF_PERSONNEL;1 Backup File = PARTIAL.RBF ! Journal = specification for journal file 2 ! Tape_Expiration = dd-mmm-yyyy ! Active_IO = number of buffers for each tape ! Protection = file system protection for backup file ! Block_Size = bytes per tape block ! Density = tape density ![No]Group_Size = number of blocks between XOR blocks ! Lock_Timeout = number of second to wait for locks ! Owner = identifier of owner of the backup file !Page_Buffers = number of buffers to use for each storage area Checksum_Verification 3 CRC = AUTODIN_II NoIncremental ! Accept_labels preserves all tape labels Log ! Loader_synchronization labels tapes in order across drives ! Media_loader forces support of a tape media loader NoOnline Quiet_Point NoRewind Statistics ACL ![No]Scan_Optimization Labels = (- 001 - 002 - 003 - 004 - 005 - 006 - 007 - 008 - 009 ) End Plan Parameters Executor Parameters : Executor Name = COORDINATOR Executor Type = Coordinator End Executor Parameters Executor Parameters : 4 Executor Name = WORKER_001 Executor Type = Worker Executor Node = NODE1 Start Storage Area List EMPIDS_LOW, SALARY_HISTORY End Storage Area List Tape Drive List Tape Drive = TAPE1: MASTER End Tape Drive List End Executor Parameters Executor Parameters : Executor Name = WORKER_002 Executor Type = Worker Executor Node = NODE2 Start Storage Area List EMPIDS_MID, RDB$SYSTEM End Storage Area List Tape Drive List Tape Drive = TAPE2: MASTER End Tape Drive List End Executor Parameters Executor Parameters : Executor Name = WORKER_003 Executor Type = Worker Executor Node = NODE1 Start Storage Area List EMPIDS_OVER End Storage Area List Tape Drive List Tape Drive = TAPE3 MASTER End Tape Drive List End Executor Parameters Executor Parameters : Executor Name = WORKER_004 Executor Type = Worker Executor Node = NODE2 Start Storage Area List EMP_INFO End Storage Area List Tape Drive List Tape Drive = TAPE4 MASTER End Tape Drive List End Executor Parameters Example 12 The following example demonstrates the use of the Restore_Options qualifier. The first command backs up selected areas of the mf_personnel database and creates an options file. The second command shows the contents of the options file. The last command demonstrates the use of the options file with the RMU Restore command. $ RMU/BACKUP MF_PERSONNEL.RDB MF_EMPIDS.RBF/INCLUDE=(EMPIDS_LOW, - _$ EMPIDS_MID, EMPIDS_OVER) /RESTORE_OPTIONS=MF_EMPIDS.OPT %RMU-I-NOTALLARE, Not all areas will be included in this backup file $ ! $ ! $ TYPE MF_EMPIDS.OPT ! Options file for database USER1:[MFDB]MF_PERSONNEL.RDB;1 ! Created 18-JUL-1995 10:31:08.82 ! Created by BACKUP command EMPIDS_LOW - /file=USER2:[STOA]EMPIDS_LOW.RDA;1 - /blocks_per_page=2 - /extension=ENABLED - /read_write - /spams - /thresholds=(70,85,95) - /snapshot=(allocation=100, - file=USER2:[SNP]EMPIDS_LOW.SNP;1) EMPIDS_MID - /file=USER3:[STOA]EMPIDS_MID.RDA;1 - /blocks_per_page=2 - /extension=ENABLED - /read_write - /spams - /thresholds=(70,85,95) - /snapshot=(allocation=100, - file=USER3:[SNP]EMPIDS_MID.SNP;1) EMPIDS_OVER - /file=USER4:[STOA]EMPIDS_OVER.RDA;1 - /blocks_per_page=2 - /extension=ENABLED - /read_write - /spams - /thresholds=(70,85,95) - /snapshot=(allocation=100, - file=USER4:[SNP]EMPIDS_OVER.SNP;1) $ ! $ ! $ ! $ RMU/RESTORE MF_EMPIDS.RBF /AREA/OPTIONS=MF_EMPIDS.OPT Example 13 The following example uses a density value with compression: $RMU/BACKUP/DENSITY=(TK89,COMPACTION)/REWIND/LABEL=(LABEL1,LABEL2) - _$ MF_PERSONNEL TAPE1:MFP.BCK, TAPE2: Example 14 The following example shows how to perform a multidisk backup operation. $ RMU/BACKUP/DISK MF_PERSONNEL DEVICE1:[DIRECTORY1]MFP.RBF, - _$ DEVICE2:[DIRECTORY2] . . . %RMU-I-COMPLETED, BACKUP operation completed at 1-MAY-2001 17:40:53.81 Example 15 The following example shows the use of the Librarian qualifier with a plan file. $RMU/BACKUP/PARALLEL=EXECUTOR=3/LIBRARIAN=WRITER_THREADS=3 - _$ /LIST_PLAN=FILENAME.PLAN/NOEXECUTE/LOG DATABASE FILENAM.RBF $RMU/BACKUP/PLAN FILENAME.PLAN $RMU/RESTORE/LIBRARIAN=(READER_THREADS=9)/LOG FILENAME.RBF The first backup command creates the plan file for a parallel backup, but does not execute it. The second backup command executes the parallel backup using the plan file. Three worker processes are used; each process uses the three writer threads specified with the Librarian qualifier. Each writer thread in each process writes one stream of backup data to the Librarian utility; a total of nine streams is written. Example 16 This example shows the use of the Compression qualifier ZLIB. $ RMU /BACKUP /COMPRESS=ZLIB:9 /LOG=FULL FOO BCK . . . BACKUP summary statistics: Data compressed by 53% (9791 KB in/4650 KB out) Example 17 The following example shows the use of the Norecord qualifier. This would be used to backup a Hot Standby database without modifying the database files. $ RMU /BACKUP /NORECORD FOO BCK 3 After_Journal Creates a backup file of the database after-image journal (.aij) file or files. Oracle Rdb supports two types of after-image journaling mechanisms: one that employs a single, extensible .aij file and another that employs multiple, fixed-size .aij files. The type of journaling mechanism being used at the time the backup operation starts can affect how you should specify the backup command. Further information on how these two journaling mechanisms affect the backup operation appears in the Description help entry under this command. The backup .aij file is an actual, usable .aij file that can be applied to the appropriate Oracle Rdb database in a recovery operation. The RMU Backup After_Journal command can be used while users are attached to the database. 4 Description The backup .aij file you create can be used with the RMU Recover command to recover (roll forward) journaled transactions. In some cases, you might have to issue additional Recover commands: one for the backup .aij file and a second for the more recent .aij files. Oracle Rdb supports the following two types of .aij file configurations: o A configuration that uses a single, extensible .aij file This is the method always used prior to Version 6.0 and is also the default (for compatibility with versions of Oracle Rdb prior to Version 6.0). When an extensible .aij file is used, one .aij file is written to and extended, as needed, by the number of blocks specified when the .aij file was created. The .aij file continues to be extended until it is backed up (or the device on which it resides is full). The RMU Backup After_Journal command copies transactions recorded in the current .aij file (always on a disk device) to the backup .aij file (which might be on a tape or disk device). On completion, the current .aij file is truncated and used again. During periods of high update activity, the truncation of the active .aij file might not be performed because of conflicting access to the .aij file by other users, but the storage allocated to the active .aij file is still used again when the backup operation completes. o A configuration that uses two or more fixed-size .aij files When fixed-size .aij files are used, the database maintains multiple .aij files; however, only one .aij file is written to at a time. This .aij file is considered the current journal. When this .aij file is filled, a switchover occurs to allow journaling to continue in another available .aij file. The RMU Backup After_Journal command works as follows with fixed-size .aij files: - Backs up any full .aij files The backup operation first backs up the .aij file with the lowest AIJ sequence number (that needs backing up), the operation continues to back up .aij files in ascending AIJ sequence number. If a lot of .aij files need to be backed up when the RMU Backup After_Journal command is issued, one backup file might contain the contents of all the .aij files being backed up. - Backs up the current .aij file Even if there are active transactions at the time of the backup operation, the RMU Backup After_Journal command will start to backup the current active .aij file. If you have specified the Quiet_Point qualifier, the backup operation stalls at some point waiting for all the current transactions to complete. - Switches to the next available .aij file An available .aij file is one for which both of the following are true: * It is not currently being used to record transactions. * It is not needed for a redo operation. Such an .aij file might be one that has never been used, or one that has already been backed up. Once a specified .aij file has been completely backed up, it is initialized and marked as available for reuse. NOTE The method employed, fixed-size .aij files or an extensible .aij file, cannot be set explicitly by the user. Any event that reduces the number of .aij files to one results in an extensible .aij file being used. Any event that increases the number .aij files to two or more results in fixed-size .aij files being used. An inaccessible .aij file is counted in these equations. Therefore, if you have one accessible .aij file and one inaccessible .aij file (perhaps because it has been suppressed), fixed-size .aij journaling is still used. Because some of the RMU Backup After_Journal qualifiers are valid only when one or the other journaling mechanism is employed, you might need to issue an RMU Dump command to determine which journaling mechanism is currently being employed before you issue an RMU Backup After_Journal command. Also note that once a backup operation begins, .aij file modification is not allowed until the backup operation is complete. However, if the type of journaling changes between the time you issue an RMU Dump command and the time you issue the RMU Backup After_Journal command, you receive an error message if you have specified qualifiers that are only valid with a particular type of journaling mechanism. (The Threshold qualifier, for example, is valid only when the extensible journaling mechanism is being used.) If you back up the .aij file or files to tape, you must mount the backup media by using the DCL MOUNT command before you issue the RMU Backup After_Journal command. If you specify the default, Format=Old_File, the RMU Backup After_Journal command uses RMS to write to the tape and the tape must be mounted as an OpenVMS volume. (That is, do not specify the FOREIGN qualifier with the MOUNT command.) If you specify the Format=New_Tape qualifier, the RMU Backup After_Journal command writes backup files in a format similar to that used by the RMU Backup command, and you must mount the tape as a FOREIGN volume. If you back up an .aij file to disk, you can then use the OpenVMS Backup utility (BACKUP) to archive the .aij backup file. The RMU Backup After_Journal command can be used in a batch job to avoid occupying an interactive terminal for long periods of time. The Continuous, Interval, Threshold, and Until qualifiers control the duration and frequency of the backup process. When you use the Continuous qualifier, the command can occupy a terminal indefinitely. Therefore, it is good practice to issue the command through a batch process when executing a continuous .aij file backup operation. However, remember that the portion of the command procedure that follows the RMU Backup After_Journal command is not executed until after the time specified by the Until qualifier. When the RMU Backup After_Journal command completes, it records information about the state of the backup files in the global process symbols presented in the following list. You can use these symbols in DCL command procedures to help automate the backup operation. These symbols are not set, however, if you have issued a DCL SET SYMBOL/SCOPE=(NOLOCAL, NOGLOBAL) command. o RDM$AIJ_SEQNO Contains the sequence number of the last .aij backup file written to tape. This symbol has a value identical to RDM$AIJ_ BACKUP_SEQNO. RDM$AIJ_SEQNO was created prior to Oracle Rdb Version 6.0 and is maintained for compatibility with earlier versions of Oracle Rdb. o RDM$AIJ_CURRENT_SEQNO Contains the sequence number of the currently active .aij file. A value of -1 indicates that after-image journaling is disabled. o RDM$AIJ_NEXT_SEQNO Contains the sequence number of the next .aij file that needs to be backed up. This symbol always contains a positive integer value (which can be 0). o RDM$AIJ_LAST_SEQNO Contains the sequence number of the last .aij file ready for a backup operation, which is different from the current sequence number if fixed-size journaling is being used. A value of -1 indicates that no journal has ever been backed up. If the value of the RDM$AIJ_NEXT_SEQNO symbol is greater than the value of the RDM$AIJ_LAST_SEQNO symbol, no more .aij files are currently available for the backup operation. o RDM$AIJ_BACKUP_SEQNO Contains the sequence number of the last .aij file backed up by the backup operation. This symbol is set at the completion of an .aij backup operation. A value of -1 indicates that this process has not yet backed up an .aij file. The RMU Backup After_Journal command provides an informational message that describes the exact sequence number for each .aij backup file operation. o RDM$AIJ_COUNT Contains the number of available .aij files. o RDM$AIJ_ENDOFFILE Contains the end of file block number for the current AIJ journal. o RDM$AIJ_FULLNESS Contains the percent fullness of the current AIJ journal. Note that these are string symbols, not integer symbols, even though their equivalence values are numbers. Therefore performing arithmetic operations with them produces unexpected results. If you need to perform arithmetic operations with these symbols, first convert the string symbol values to numeric symbol values using the OpenVMS F$INTEGER lexical function. For example: $ SEQNO_RANGE = F$INTEGER(RDB$AIJ_LAST_SEQNO) - F$INTEGER(RDB$AIJ_NEXT_SEQNO) 4 Format (B)0 RMU/Backup/After_Journal root-file-spec {backup-file-spec | ""} Command Qualifiers x Defaults /[No]Accept_Label x /Accept_Label /Active_IO=max-writes x /Active_IO=3 /Block_Size=integer x See description /[No]Compression[=options] x /Nocompression /[No]Continuous=(n) x /Nocontinuous /[No]Crc x See description /Crc[=Autodin_II] x See description /Crc=Checksum x See description /Density=(density-value, [No]Compaction) x See description /[No]Edit_Filename=(options) x /Noedit_Filename /Encrypt=({Value=|Name=}[,Algorithm=]) x See description /Format={Old_File|New_Tape} x /Format=Old_File /[No]Group_Size[=interval] x See description /[No]Interval=number-seconds x /Nointerval /Label=(label-name-list) x See description /Librarian[=options] x None /Lock_Timeout=seconds x See description /[No]Log x Current DCL verify value (B)0 /[No]Media_Loader x See description /Owner=user-id x See description /Prompt={Automatic|Operator|Client} x See description /Protection=openvms-file-protection x See description /[No]Quiet_Point x /Quiet_Point /[No]Rename x /Norename /[No]Rewind x /Norewind /[No]Sequence=(n,m) x /Nosequence /Tape_Expiration=date-time x The current time /[No]Threshold=disk-blocks x /Nothreshold /Until=time x See description /[No]Wait=n x See description 4 Parameters 5 root-file-spec The name of the database root file. The root file name is also the name of the database. An error results if you specify a database that does not have after-image journaling enabled. The default file extension is .rdb. 5 backup-file-spec A file specification for the .aij backup file. The default file extension is .aij unless you specify the Format=New_Tape qualifier. In this case, the default file extension is .aij_rbf. 5 "" Double quotes indicate to Oracle RMU that you want the default .aij backup file specification to be used. The default .aij backup file specification is defined with the SQL ALTER DATABASE statement or the RMU Set After_Journal command. 4 Command_Qualifiers 5 Accept_Label Accept_Label Specifies that Oracle RMU should keep the current tape label it finds on a tape during a backup operation even if that label does not match the default label or that specified with the Label qualifier. Operator notification does not occur unless the tape's protection, owner, or expiration date prohibit writing to the tape. However, a message is logged (assuming logging is enabled) and written to the backup journal file (assuming you have specified the Journal qualifier) to indicate that a label is being preserved and which drive currently holds that tape. This qualifier is particularly useful when your backup operation employs numerous previously used (and thus labeled) tapes and you want to preserve the labels currently on the tapes. If you do not specify this qualifier, the default behavior of Oracle RMU is to notify the operator each time it finds a mismatch between the current label on the tape and the default label (or the label you specify with the Label qualifier). See the description of the Labels qualifier under this command for information on default labels. See How Tapes are Relabeled During a Backup Operation in the Usage_Notes help entry under the Backup Database help entry for a summary of which labels are applied under a variety of circumstances. 5 Active_IO Active_IO=max-writes Specifies the maximum number of write operations to a backup device that the RMU Backup After_Journal command attempts simultaneously. This is not the maximum number of write operations in progress; that value is the product of active system I/O operations and the number of devices being written to simultaneously. The value of the Active_IO qualifier can range from 1 to 5. The default value is 3. Values larger than 3 can improve performance with some tape drives. 5 Block_Size Block_Size=integer Specifies the maximum record size for the backup file. The size can vary between 2048 and 65,024 bytes. The default value is device dependent. The appropriate block size is a compromise between tape capacity and error rate. 5 Compression Compression=LZSS Compression=Huffman Compression=ZLIB=level Nocompression Allows you to specify the compression method to use before writing data to the AIJ backup file. This reduces performance, but may be justified when the AIJ backup file is a disk file, or is being backed up over a busy network, or is being backed up to a tape drive that does not do its own compression. You probably do not want to specify the Compression qualifier when you are backing up an aIJ file to a tape drive that does its own compression; in some cases doing so can actually result in a larger file. This feature only works for the new backup file format and you have to specify /FORMAT=NEW_TAPE if you also use /COMPRESSION. If you specify the Compression qualifier without a value, the default is COMPRESSION=ZLIB=6. The level value (ZLIB=level) is an integer between 1 and 9 specifying the relative compression level with one being the least amount of compression and nine being the greatest amount of compression. Higher levels of the compression use increased CPU time while generally providing better compression. The default compression level of 6 is a balance between compression effectiveness and CPU consumption. OLDER ORACLE RDB 7.2 RELEASES AND COMPRESSED RBF FILES Prior releases of Oracle Rdb are unable to read RBF files compressed with the ZLIB algorithm. In order to read compressed backups with Oracle Rdb 7.2 Releases prior to V7.2.1, they must be made with /COMPRESSION=LZSS or /COMPRESSION=HUFFMAN explicitly specified (because the default compression algorithm has been changed from LZSS to ZLIB). Oracle Rdb Version 7.2.1 is able to read compressed backups using the LZSS or HUFFMAN algorithms made with prior releases. 5 Continuous Continuous=(n) Nocontinuous Specifies whether the .aij backup process operates continuously. You specify termination conditions by specifying one or both of the following: o The Until qualifier Specifies the time and date to stop the continuous backup process. o The value for n Specifies the number of iterations Oracle RMU should make through the set of active .aij files before terminating the backup operation. When you use the Continuous qualifier, you must use either the Until or the Interval qualifier or provide a value for n (or both) to specify when the backup process should stop. You can also stop the backup process by using the DCL STOP command when backing up to disk. If you specify the Continuous qualifier, Oracle Rdb does not terminate the backup process after truncating the current .aij file (when an extensible journal is used) or after switching to a new journal (when fixed-size journals are used). Instead, the backup process waits for the period of time that you specify in the argument to the Interval qualifier. After that time interval, the backup process tests to determine if the threshold has been reached (for an extensible journal) or if the journal is full (for fixed-size journals). It then performs backup operations as needed and then waits again until the next interval break, unless the number of iterations or the condition specified with the Until qualifier has been reached. If you specify the Continuous qualifier, the backup process occupies the terminal (that is, no system prompt occurs) until the process terminates. Therefore, you should usually enter the command through a batch process. If you specify the default, the Nocontinuous qualifier, the backup process stops as soon as it completely backs up the .aij file or files. The default value for the number of iterations (n) is 1. If you specify both the Until qualifier and the Continuous=n qualifier, the backup operation stops after whichever completes first. If you specify the Until=12:00 qualifier and the Continuous=5 qualifier, the backup operation terminates at 12:00 even if only four iterations have completed. Likewise, if five iterations are completed prior to 12:00, the backup operation terminates after the five iterations are completed. The Continuous qualifier is not recommended when you are backing up to tape, particularly when the Format=New_Tape qualifier is used. If your tape operations complete successfully, you do not want the backup operation to continue in an infinite loop. Using the DCL STOP command to terminate a backup operation to tape might result in an incomplete or corrupt backup file. However, do not delete this backup file; it is extremely important that you preserve all .aij backup files, even those produced by failed or terminated backup processes. If the resultant .aij backup file is discarded, the next .aij backup file could contain a "gap" in transactions, so that no transactions would ever be rolled forward from that point on. 5 Crc[=Autodin_II] Crc[=Autodin_II] Uses the AUTODIN-II polynomial for the 32-bit CRC calculation and provides the most reliable end-to-end error detection. This is the default for NRZ/PE (800/1600 bits/inch) tape drives. Typing Crc is sufficient to select the Crc=Autodin_II qualifier. It is not necessary to type the entire qualifier. 5 Crc=Checksum Crc=Checksum Uses one's complement addition, which is the same computation used to do a checksum of the database pages on disk. This is the default for GCR (6250 bits/inch) tape drives and for TA78, TA79, and TA81 tape drives. The Crc=Checksum qualifier allows detection of errors. 5 Nocrc Nocrc Disables end-to-end error detection. This is the default for TA90 (IBM 3480 class) drives. NOTE The overall effect of the Crc=Autodin_II, Crc=Checksum, and Nocrc qualifier defaults is to improve tape reliability so that it is equal to that of a disk. If you retain your tapes longer than 1 year, the Nocrc default might not be adequate. For tapes retained longer than 1 year, use the Crc=Checksum qualifier. If you retain your tapes longer than 3 years, you should always use the Crc=Autodin_II qualifier. Tapes retained longer than 5 years could be deteriorating and should be copied to fresh media. See the Oracle Rdb Guide to Database Maintenance for details on using the Crc qualifiers to avoid underrun errors. 5 Density Density=(density-value,[No]Compaction) Specifies the density at which the output volume is to be written. The default value is the format of the first volume (the first tape you mount). You do not need to specify this qualifier unless your tape drives support data compression or more than one recording density. The Density qualifier is applicable only to tape drives. Oracle RMU returns an error message if this qualifier is used and the target device is not a tape drive. If your systems are running OpenVMS versions prior to 7.2-1, specify the Density qualifier as follows: o For TA90E, TA91, and TA92 tape drives, specify the number in bits per inch as follows: - Density = 70000 to initialize and write tapes in the compacted format. - Density = 39872 or Density = 40000 for the noncompacted format. o For SCSI (Small Computer System Interface) tape drives, specify Density = 1 to initialize and write tapes using the drive's hardware data compression scheme. o For other types of tape drives, you can specify a supported Density value between 800 and 160000 bits per inch. o For all tape drives, specify Density = 0 to initialize and write tapes at the drive's standard density. Do not use the Compaction or NoCompaction keyword for systems running OpenVMS versions prior to 7.2-1. On these systems, compression is determined by the density value and cannot be specified. Oracle RMU supports the OpenVMS tape density and compression values introduced in OpenVMS Version 7.2-1. The following table lists the added density values supported by Oracle RMU. DEFAULT 800 833 1600 6250 3480 3490E TK50 TK70 TK85 TK86 TK87 TK88 TK89 QIC 8200 8500 8900 DLT8000 SDLT SDLT320 SDLT600 DDS1 DDS2 DDS3 DDS4 AIT1 AIT2 AIT3 AIT4 LTO2 LTO3 COMPACTION NOCOMPACTION If the OpenVMS Version 7.2-1 density values and the previous density values are the same (for example, 800, 833, 1600, 6250), the specified value is interpreted as an OpenVMS Version 7.2-1 value if the tape device driver accepts them, and as a previous value if the tape device driver accepts previous values only. For the OpenVMS Version 7.2-1 values that accept tape compression you can use the following syntax: /DENSITY = (new_density_value,[No]Compaction) In order to use the Compaction or NoCompaction keyword, you must use one of the following density values that accepts compression: DEFAULT 3480 3490E 8200 8500 8900 TK87 TK88 TK89 DLT8000 SDLT SDLT320 AIT1 AIT2 AIT3 AIT4 DDS1 DDS2 DDS3 DDS4 SDLT600 LTO2 LTO3 Refer to the OpenVMS documentation for more information about density values. 5 Edit_Filename Edit_Filename=(Options) Noedit_Filename When the Edit_Filename=(options) qualifier is used, the specified backup file name is edited by appending any or all of the values specified by the following options to the backup file name: o Day_Of_Week The current day of the week expressed as a 1-digit integer (1 to 7). Sunday is expressed as 1; Saturday is expressed as 7. o Day_Of_Year The current day of the year expressed as a 3-digit integer (001 to 366). o Hour The current hour of the day expressed as a 2-digit integer (00 to 23). o Julian_Date The number of days passed since 17-Nov-1858. o Minute The current minute of the hour expressed as a 2-digit integer (00 to 59). o Month The current month expressed as a 2-digit integer (01 to 12). o Sequence The journal sequence number of the first journal in the backup operation. o Vno Synonymous with the Sequence option. See the description of the Sequence option. o Year The current year (A.D.) expressed as a 4-digit integer. If you specify more than one option, place a comma between each option. The edit is performed in the order specified. For example, the file backup.aij and the qualifier /EDIT_FILENAME=(HOUR, MINUTE, MONTH, DAY_OF_MONTH, SEQUENCE) creates a file with the name backup_160504233.aij when journal 3 is backed up at 4:05 P.M. on April 23rd. You can make the name more readable by inserting quoted strings between each Edit_Filename option. For example, the following qualifier adds the string "$30_0155-2" to the .aij file name if the day of the month is the 30th, the time is 1:55 and the version number is 2: /EDIT_FILENAME=("$",DAY_OF_MONTH,"_",HOUR,MINUTE,"-",SEQUENCE) This qualifier is useful for creating meaningful file names for your backup files and makes file management easier. The default is the Noedit_Filename qualifier. 5 Encrypt Encrypt=({Value=|Name=}[,Algorithm=]) The Encrypt qualifier encrypts the backup file of the after image journal. Specify a key value as a string or, the name of a predefined key. If no algorithm name is specified the default is DESCBC. For details on the Value, Name and Algorithm parameters see HELP ENCRYPT. This feature requires the OpenVMS Encrypt product to be installed and licensed on this system. This feature only works for a newer format backup file which has been created using /FORMAT=NEW_TAPE. Therefore you have to specify /FORMAT=NEW_TAPE with this command if you also use /ENCRYPT. Synonymous with Format=Old_File and Format=New_Tape qualifiers. See the description of those qualifiers. 5 Format Format=Old_File Format=New_Tape Specifies the format in which the backup file is to be written. Oracle Corporation recommends that you specify the Format=Old_ File qualifier (or accept the default) when you back up your .aij file to disk and that you specify the Format=New_Tape qualifier when you back up your .aij file to tape. If you specify the default, the Format=Old_File qualifier, the RMU Backup command writes the file in a format that is optimized for a file structured disk. If you specify the Format=New_Tape qualifier, the Oracle RMU command writes the file in a format that is optimized for tape storage, including ANSI/ISO labeling and end-to-end error detection and correction. When you specify the Format=New_Tape qualifier and back up the .aij file to tape, you must mount the backup media by using the DCL MOUNT command before you issue the RMU Backup After_Journal command. The tape must be mounted as a FOREIGN volume. If you mount the tape as an OpenVMS volume (that is, you do not mount it as a FOREIGN volume) and you specify the Format=New_Tape qualifier, you receive an RMU-F-MOUNTFOR error. When you back up your .aij file to tape and specify the Format=New_Tape qualifier you can create a backup copy of the database (using the RMU Backup command) and a backup of the .aij file (using the RMU Backup After_Journal command) without dismounting your tape. The following tape qualifiers have meaning only when used in conjunction with the Format=New_Tape qualifier: Active_IO Block_Size Crc Density Group_Size Label Owner Protection Rewind Tape_Expiration The Format=New_Tape and the Noquiet_Point qualifiers cannot be used on the same Oracle RMU command line. See the Usage Notes Help entry for an explanation. The default file specification, when you specify the Format=New_ Tape qualifier is .aij_rbf. The default file specification, when you specify the Format=Old_File qualifier is .aij. Although Oracle Corporation recommends that you specify the Format=New_Tape qualifier for .aij backup operations to tape and the Format=Old_File qualifier for .aij backup operations to disk, Oracle RMU does not enforce this recommendation. This is to provide compatibility with prior versions of Oracle Rdb. See the Usage Notes Help entry for issues and problems you can encounter when you do not follow this recommendation. 5 Group_Size Group_Size[=interval] Nogroup_Size Specifies the frequency at which XOR recovery blocks are written to tape. The group size can vary from 0 to 100. Specifying a group size of zero or specifying the Nogroup_Size qualifier results in no XOR recovery blocks being written. The Group_Size qualifier is only applicable to tape, and its default value is device dependent. Oracle RMU returns an error message if this qualifier is used and the target device is not a tape device. 5 Interval=n Interval=number-seconds Nointerval Specifies the number of seconds for which the backup process waits. Use this qualifier in conjunction with the Continuous qualifier and the extensible journaling method. The interval determines how often to test the active .aij file to determine if it contains more blocks than the value of the Threshold qualifier. If you specify the Interval qualifier without specifying the number of seconds, or if you omit this qualifier, the default number of seconds is 60. Oracle Corporation recommends using the default (Interval=60) initially because the interval that you choose can affect the performance of the database. In general, you can arrive at a good interval time on a given database only by judgment and experimentation. If you specify the Nointerval qualifier, the active .aij file is tested repeatedly with no interval between finishing one cycle and beginning the next. You must specify the Continuous qualifier if you specify either the Interval or Nointerval qualifier. If you specify both the Interval and Nocontinuous qualifiers, the Interval qualifier is ignored. 5 Label Label=(label-name-list) Specifies the 1- to 6-character string with which the volumes of the backup file are to be labeled. The Label qualifier is applicable only to tape volumes. You must specify one or more label names when you use the Label qualifier. You can specify a list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas and enclose the list of names within parentheses. If you do not specify the Label (or Accept_Label) qualifier, Oracle RMU labels the first tape used for a backup operation with the first 6 characters of the backup file name. Subsequent default labels are the first 4 characters of the backup file name appended with a sequential number. For example, if your backup file is my_backup.rbf, the default tape labels are my_b, my_b01, my_b02, and so on. When you reuse tapes, Oracle RMU compares the label currently on the tape to the label or labels you specify with the Label qualifier. If there is a mismatch between the existing label and a label you specify, Oracle RMU sends a message to the operator asking if the mismatch is acceptable (unless you also specify the Accept_Labels qualifier). If desired, you can explicitly specify the list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas and enclose the list of names within parentheses. If you are reusing tapes be certain that you load the tapes so that the label Oracle RMU expects and the label on each tape will match, or be prepared for a high level of operator intervention. If you specify fewer labels than are needed, Oracle RMU generates labels based on the format you have specified. For example, if you specify Label=TAPE01, Oracle RMU labels subsequent tapes as TAPE02, TAPE03, and so on up to TAPE99. Thus, many volumes can be preloaded in the cartridge stacker of a tape drive. The order is not important because Oracle RMU relabels the volumes. An unattended backup operation is more likely to be successful if all the tapes used do not have to be mounted in a specific order. Once the backup operation is complete, externally mark the tapes with the appropriate label so that the order can be maintained for the restore operation. Be particularly careful if you are allowing Oracle RMU to implicitly label second and subsequent tapes and you are performing an unattended backup operation. Remove the tapes from the drives in the order in which they were written. Apply labels to the volumes following the logic of implicit labeling (for example, TAPE02, TAPE03, and so on). Oracle Corporation recommends you use the Journal qualifier when you employ implicit labeling in a multidrive, unattended backup operation. The journal file records the volume labels that were written to each tape drive. The order in which the labels were written is preserved in the journal. Use the RMU Dump Backup command to display a listing of the volumes written by each tape drive. You can use an indirect file reference with the Label qualifier. See the Indirect-command-files help entry for more information. See How Tapes are Relabeled During a Backup Operation in the Usage_Notes help entry under this command for a summary of which labels are applied under a variety of circumstances. 5 Librarian Librarian=options Use the Librarian qualifier to back up files to data archiving software applications that support the Oracle Media Management interface. The backup file name specified on the command line identifies the stream of data to be stored in the Librarian utility. If you supply a device specification or a version number it will be ignored. The Librarian qualifier accepts the following options: o Trace_file=file-specification The Librarian utility writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian utility. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian utility. The pre-determined values are : - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical_name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian utility can use to specify catalogs or archives where Oracle Rdb backup files are stored, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of any Oracle RMU command that accesses the Librarian utility. The following OpenVMS logical names must be defined for use with a Librarian utility before you execute an Oracle RMU backup or restore operation. Do not use the Logical_Names option provided with the Librarian qualifier to define these logical names. o RMU$LIBRARIAN_PATH This logical name must be defined so that the shareable Librarian image can be loaded and called by Oracle RMU backup and restore operations. The translation must include the file type (for example, .exe), and must not include a version number. The shareable Librarian image must be an installed (known) image. See the Librarian utility documentation for the name and location of this image and how it should be installed. o RMU$DEBUG_SBT This logical name is not required. If it is defined, Oracle RMU will display debug tracing information messages from modules that make calls to the Librarian shareable image. You cannot use device specific qualifiers such as Rewind, Density, or Label with the Librarian qualifier because the Librarian utility handles the storage meda, not Oracle RMU. 5 Lock_Timeout Lock_Timeout=seconds Determines the maximum time the .aij file backup operation will wait for the quiet-point lock and any other locks needed during online backup operations. When you specify the Lock_ Timeout=seconds qualifier, you must specify the number of seconds to wait for the quiet-point lock. If the time limit expires, an error is signaled and the backup operation fails. When the Lock_Timeout=seconds qualifier is not specified, or if the value specified is 0, the .aij file backup operation waits indefinitely for the quiet-point lock and any other locks needed during an online operation. The Lock_Timeout=seconds qualifier is ignored if the Noquiet_ Point qualifier is specified. 5 Log Log Nolog Specifies whether the processing of the command is reported to SYS$OUTPUT. Specify the Log qualifier to request log output and the Nolog qualifier to prevent it. If you specify neither, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 5 Media_Loader Media_Loader Nomedia_Loader Use the Media_Loader qualifier to specify that the tape device receiving the backup file has a loader or stacker. Use the Nomedia_Loader qualifier to specify that the tape device does not have a loader or stacker. By default, if a tape device has a loader or stacker, Oracle RMU should recognize this fact. However, occasionally Oracle RMU does not recognize that a tape device has a loader or stacker. Therefore, when the first backup tape fills, Oracle RMU issues a request to the operator for the next tape, instead of requesting the next tape from the loader or stacker. Similarly, sometimes Oracle RMU behaves as though a tape device has a loader or stacker when actually it does not. If you find that Oracle RMU is not recognizing that your tape device has a loader or stacker, specify the Media_Loader qualifier. If you find that Oracle RMU expects a loader or stacker when it should not, specify the Nomedia_Loader qualifier. Synonymous with Owner qualifier. See the description of the Owner qualifier. 5 Owner Owner=user-id Specifies the owner of the tape volume set. The owner is the user who will be permitted to restore the database. The user- id parameter must be one of the following types of OpenVMS identifier: o A user identification code (UIC) in [group-name,member-name] alphanumeric format o A UIC in [group-number,member-number] numeric format o A general identifier, such as SECRETARIES o A system-defined identifier, such as DIALUP The Owner qualifier cannot be used with a backup operation to disk. When used with tapes, the Owner qualifier applies to all continuation volumes. Unless the Rewind qualifier is also specified, the Owner qualifier is not applied to the first volume. If the Rewind qualifier is not specified, the backup operation appends the file to a previously labeled tape, so the first volume can have a protection different from the continuation volumes. 5 Prompt Prompt=Automatic Prompt=Operator Prompt=Client Specifies where server prompts are to be sent. When you specify Prompt=Automatic, prompts are sent to the standard input device, and when you specify Prompt=Operator, prompts are sent to the server console. When you specify Prompt=Client, prompts are sent to the client system. 5 Protection Protection=file-protection Specifies the system file protection for the backup file produced by the RMU Backup After_Journal command. The default file protection varies, depending on whether you backup the file to disk or tape. This is because tapes do not allow delete or execute access and the SYSTEM account always has both read and write access to tapes. In addition, a more restrictive class accumulates the access rights of the less restrictive classes. If you do not specify the Protection qualifier, the default protection is as follows: o S:RWED,O:RE,G,W if the backup is to disk o S:RW,O:R,G,W if the backup is to tape If you specify the Protection qualifier explicitly, the differences in protection applied for backups to tape or disk as noted in the preceding paragraph are applied. Thus, if you specify Protection=(S,O,G:W,W:R), that protection on tape becomes (S:RW,O:RW,G:RW,W:R). 5 Quiet_Point Quiet_Point Noquiet_Point Specifies whether the quiet-point lock will be acquired when an .aij backup operation is performed. The default is the Quiet_ Point qualifier. Use of the Quiet_Point qualifier is meaningful only for a full backup operation; that is, a backup operation that makes a complete pass through all .aij files ready for backup as opposed to one which is done by-sequence (specified with the Sequence qualifier). A full .aij backup operation can be performed regardless of whether an extensible or a fixed-size .aij journaling mechanism is being employed. Each .aij backup operation is assigned an .aij sequence number. This labeling distinguishes each .aij backup file from previous .aij backup files. During a recovery operation, it is important to apply the .aij backup files in the proper sequence. The RMU Recover command checks the database root file structure and displays a message telling you the .aij sequence number with which to begin the recovery operation. The quiet point is a state where all write transactions have either been committed or rolled back and no read/write transactions are in progress. This ensures that the recording of transactions do not extend into a subsequent .aij backup file. This backup file can then be used to produce a recovered database that is in the same state as when the quiet point was reached. When fixed-size journaling is employed, the Quiet_Point qualifier is only relevant when the active .aij file is being backed up. In this case, a quiet point is acquired only once, regardless of the number of .aij files being backed up. There is no natural quiet point if someone is writing or waiting to write to the database at any given time. (A natural quiet point is one that is not instigated by the use of the QP (quiet point) Lock.) The .aij backup operation may never be able to capture a state that does not have uncommitted data in the database. As a result, the Noquiet_Point qualifier creates .aij backup files that are not independent of one another. If you apply one .aij backup file to the database without applying the next .aij backup file in sequence, the recovery operation will not be applied completely. See the Usage_Notes help entry under this command for recommendations on using the Quiet_Point and Noquiet_Point qualifiers. The following combination of qualifiers on the same command line are invalid: o Quiet_Point and Sequence o Quiet_Point and Wait o Noquiet_Point and Format=New 5 Rename Rename Norename The Rename qualifier creates and initializes a new .aij file and creates the backup file by renaming the original .aij file. The effect is that the original .aij file has a new name and the new .aij file has the same name as the original .aij file. The Rename qualifier sets the protection on the renamed backup file so that you can work with it as you would any backup file. You can specify the new name by using the Edit_Filename qualifier. When the Rename qualifier is used, the backup operation is faster (than when Norename, the default, is specified) because the duration of the backup operation is the total time required to rename and initialize the .aij file; the data copy portion of the backup (reading and writing) is eliminated. However, the disk containing the .aij file must have sufficient space for both the new and original .aij files. Note also that the .aij backup file name must not include a device specification. NOTE If there is insufficient space for both the new and original .aij files when the Rename qualifier is specified, after- image journaling shutdown is invoked, resulting in a complete database shutdown. The Rename qualifier can be used with both fixed-size and extensible journaling files. The Norename qualifier copies the contents of the .aij file on tape or disk and initializes the original .aij file for reuse. The Norename qualifier results in a slower backup operation (than when Rename is specified), but it does not require space on the journal disk for both new and original .aij files. The default is Norename. 5 Rewind Rewind Norewind Specifies that the magnetic tape that contains the backup file will be rewound before processing begins. The tape is initialized according to the Label and Density qualifiers. The Norewind qualifier is the default and causes the backup file to be created starting at the current logical end-of-tape (EOT). These qualifiers are applicable only to tape devices. 5 Sequence Sequence=(n,m) Nosequence Specifies that the journals with sequence numbers from n to m inclusive are to be backed up. The values n and m are interpreted or interpolated as follows: o If Sequence = (33, 35) is specified, then the .aij files with sequence numbers 33, 34, and 35 are backed up. o If Sequence = (53, 53) is specified, then the .aij file with sequence number 53 is backed up. o If Sequence = (53) is specified, then the .aij files with sequence numbers 53 and lower are backed up, if they have not been backed up already. For example, if .aij files with sequence numbers 51, 52, and 53 have not been backed up, then Sequence = (53) results in these three .aij files being backed up. o If Sequence = (55, 53) is specified, then .aij files with sequence numbers 53, 54, and 55 are backed up. o If the Sequence qualifier is specified without a value list, both n and m are set to the sequence number of the next journal that needs to be backed up. The default is the Nosequence qualifier. When the default is accepted, the backup operation starts with the next journal that needs to be backed up and stops when the termination condition you have specified is reached. The following qualifiers cannot be used or have no effect when used with the Sequence qualifier: Continuous Format=New_Tape Interval Quiet_Point Threshold Until Furthermore, fixed-size after-image journals must be in use when this qualifier is specified. 5 Tape_Expiration Tape_Expiration=date-time Specifies the expiration date of the .aij backup file. Note that when Oracle RMU reads a tape, it looks at the expiration date in the file header of the first file on the tape and assumes the date it finds in that file header is the expiration date for the entire tape. Therefore, if you are backing up an .aij file to tape, specifying the Tape_Expiration qualifier only has meaning if the .aij file is the first file on the tape. You can guarantee that the .aij file will be the first file on the tape by specifying the Rewind qualifier and overwriting any existing files on the tape. When the first file on the tape contains an expiration date in the file header, you cannot overwrite the tape before the expiration date unless you have the OpenVMS SYSPRV or BYPASS privilege. Similarly, when you attempt to perform a recover operation with an .aij file on tape, you cannot perform the recover operation after the expiration date recorded in the first file on the tape unless you have the OpenVMS SYSPRV or BYPASS privilege By default, no expiration date is written to the .aij file header. In this case, if the .aij file is the first file on the tape, the tape can be overwritten immediately. If the .aij file is not the first file on the tape, the ability to overwrite the tape is determined by the expiration date in the file header of the first file on the tape. You cannot explicitly set a tape expiration date for an entire volume. The volume expiration date is always determined by the expiration date of the first file on the tape. The Tape_ Expiration qualifier cannot be used with a backup operation to disk. See the Oracle Rdb Guide to Database Maintenance for information on tape label processing. 5 Threshold Threshold=disk-blocks Nothreshold This qualifier can be used only when extensible journaling is enabled. It cannot be used with fixed-size journaling. The Threshold qualifier sets an approximate limit on the size of the active .aij file. When the size of the active .aij file exceeds the threshold, you cannot initiate new transactions until the backup process finishes backing up and truncating (resetting) the active .aij file. During the backup operation, existing transactions can continue to write to the .aij file. Before new transactions can start, all activity issuing from existing transactions (including activity occurring after the threshold is exceeded) must be moved from the active .aij disk file to the .aij backup file. At that time, the active .aij file will be completely truncated. If you use the default, the Nothreshold qualifier, each backup cycle will completely back up the active .aij file. Oracle Corporation recommends using the Nothreshold qualifier. An appropriate value for the Threshold qualifier depends on the activity of your database, how much disk space you want to use, whether backup operations will be continuous, and how long you are willing to wait for a backup operation to complete. See the Oracle Rdb7 Guide to Database Performance and Tuning for more information on setting SPAM thresholds. 5 Until Until=time Specifies the approximate future time and date to stop the continuous backup process. There is no default. 5 Wait Wait=n Nowait Specifies whether the backup operation should wait (the Wait qualifier) or terminate (the Nowait qualifier) when it encounters a journal that is not ready to be backed up. The value specified for the Wait qualifier is the time interval in seconds between attempts to back up the journal that was not ready. The Wait or Nowait qualifier can only be specified if the Sequence qualifier is also specified. When the Wait qualifier is specified, the default value for the time interval is 60 seconds. The default is the Nowait qualifier. 4 Usage_Notes o To use the RMU Backup After_Journal command for a database, you must have the RMU$BACKUP privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o See the Oracle Rdb7 Guide to Database Performance and Tuning for information on how to enhance the performance of the RMU Backup After_Journal command. NOTE When fast commit is enabled and an extensible .aij file configuration is used, the after-image journal backup process compresses and retains some fraction of the original .aij file (in a new version of the current .aij file). This fraction can approach 100% of the original size. Therefore, be sure to reserve enough space to duplicate the maximum size .aij file before backing it up. Oracle Corporation recommends that you schedule .aij backup operations with sufficient frequency and check the free space and journal file size periodically; you need to know when you are approaching a critical situation in terms of free space. (This is good practice whether or not you have fast commit enabled.) However, if you issue the RMU Backup After_Journal command with fast commit enabled and find that you have insufficient space for the .aij file, you have the following options: o Delete unneeded files to create sufficient space on the disk where the .aij file is located. o Temporarily disable fast commit and back up the .aij file. o Close the database, disable after-image journaling, enable a new after-image journal file, and perform a backup operation. (The database can be opened either before or after the backup operation.) o Close the database. Create a bound volume set or stripe set that is large enough for the .aij file and copy the .aij file there. Use the RMU Set After_ Journal command to change the .aij file name (or redefine the logical name if one was used to locate the journal), and then open the database again. o Note the following issues and problems you can encounter when you specify the Format=Old_File qualifier for an .aij backup operation to tape or the Format=New_Tape qualifier for an .aij backup operation to disk: - If you use the Format=Old_File qualifier for an .aij backup operation to tape and the tape is mounted as a FOREIGN volume, the result is an unlabeled tape that can be difficult to use for recovery operations. Therefore, if you use the Format=Old_File qualifier with an .aij backup operation to tape, you must mount the tape as an OpenVMS volume (that is, do not specify the /FOREIGN qualifier with the DCL MOUNT command). - You must remember (or record) the format you use when you back up your .aij file and specify that same format when you issue an RMU Dump After_Journal, RMU Optimize After_ Journal, or RMU Recover command for the .aij backup file. If you always follow the guidelines of specifying Format=New_Tape for tape backups and Format=Old_File for disk backups, you do not need to track the format you specified for the .aij backup operation for future use with the other Oracle RMU .aij commands. - If you specify Format=Old_File for a backup operation to tape and the .aij spans tape volumes, you might have problems recovering the .aij file. o You can use the RMU Backup After_Journal command to save disk space by spooling the .aij file to tape. o When you use extensible .aij files, note that although a new version of the .aij file might be created when the after-image backup operation begins, the old .aij file continues to be active and growing. Until the switch occurs (which could be several hours after the creation of the new version of the .aij file), the old .aij file is still being accessed. For this and other reasons, you should never use the DCL DELETE or DCL PURGE on .aij files (or any database files). o The following list provides usage information for the Quiet_ Point and Noquiet_Point qualifiers: - If the backup operation stalls when you attempt a quiet- point Oracle RMU backup operation, it may be because another user is holding the quiet-point lock. In some cases, there is no way to avoid this stall. However, you may find the stall is caused by a user who has previously issued and completed a read-write transaction, and is currently running a read-only transaction. When this user started the read-write transaction his or her process acquired the quiet-point lock. Ordinarily, such a process retains this lock until it detaches from the database. You can set the RDM$BIND_SNAP_QUIET_POINT logical name to control whether or not such a process retains the quiet- point lock. Set the value of the logical name to "1" to allow such a process to hold the quiet-point lock until they detach from the database. Set the value of the logical name to "0", to ensure that the process releases the quiet- point lock prior to starting a read-only transaction. - When devising your backup strategy for both the database and the after-image journal files, keep in mind the trade- offs between performing quiet-point backup operations and noquiet-point backup operations. A noquiet-point backup operation is quicker than a quiet-point backup operation, but usually results in a longer recovery operation. Because transactions can span .aij files when you perform noquiet- point .aij backup operations, you might have to apply numerous .aij files to recover the database. In a worst- case scenario, this could extend back to your last quiet- point .aij or database backup operation. If you rarely perform quiet-point backup operations, recovery time could be excessive. One method you can use to balance these trade-offs is to perform regularly scheduled quiet-point .aij backup operations followed by noquiet-point database backup operations. (You could do the converse, but a quiet- point backup of the .aij file improves the performance of the recovery operation should such an operation become necessary.) Periodically performing a quiet-point .aij backup operation helps to ensure that your recovery time will not be excessive. - You cannot specify the Noquiet_Point qualifier with the Format=New_Tape qualifier because an .aij file created with the Noquiet_Point qualifier does not end on a quiet point. Some transactions can bridge several backup files. When you recover from these backup files you frequently must apply several backup files in the same RMU Recover command. However, the RMU Recover command with the Format=New_Tape qualifier can only process one backup file at a time, so it cannot support backup files created with the Noquiet_Point qualifier. o Oracle RMU tape operations do not automatically allocate the tape drives used. In an environment where many users compete for a few tape drives, it is possible for another user to seize a drive while Oracle RMU is waiting for you to load the next tape volume. To prevent this, issue a DCL ALLOCATE command for the drives you will be using before you issue the Oracle RMU command, and then issue a DCL DEALLOCATE command after you complete the Oracle RMU command. o The Label qualifier can be used with indirect file reference. See the Indirect-Command-Files help entry for more information. o If an .aij backup process fails or is terminated prematurely, the user might discard the resultant .aij backup file because the backup operation was not completed. However, all .aij backup files, including those produced by a failed backup process, are necessary to recover a database. If an .aij backup file of a failed backup process is discarded, the database is not recoverable from that point forward. This is especially important if you use magnetic tapes as the .aij backup media; in this case, preserve this magnetic tape and do not reuse it. o When an .aij backup process, especially one running in continuous (Continuous) mode, writes to the .aij backup file, it is possible for the transferred data to be deleted from the database .aij file. If the backup process subsequently fails or is prematurely terminated (for example with Ctrl/Y or the DCL STOP command), it might not be possible to retransfer the data to the subsequent .aij backup file because the data was deleted from the active database .aij file. Therefore, it is extremely important that you preserve all .aij backup files, even those produced by failed or terminated backup processes. If the resultant .aij backup file is discarded, the next .aij backup file could contain a "gap" in transactions, so that no transactions would ever be rolled forward from that point on. This problem is more severe when backing up directly to tape. Therefore, when backing up to tape, you should back up one journal at a time, rather than using an open-ended or long- duration backup operation. NOTE If this problem occurs, the database is not inconsistent or corrupt. Rather, the database cannot be rolled forward past the discarded .aij backup file. The solution to this problem is to preserve all .aij backup files to ensure that a database can be completely recovered. If you have discarded an .aij backup file, perform a full and complete database backup operation immediately to ensure that the database can be restored up to the current transaction. o When an AIJ backup operation completes, the after-image journal files are initialized with a pattern of -1 (hex FF) bytes. This initialization is designed to be as fast as possible. It fully utilizes the I/O subsystem by performing many large asynchronous I/O operations at once. However, this speed can come at the cost of a high load on I/O components during the initialization. This load could slow down other I/O operations on the system. You can use two logical names to control the relative I/O load that the AIJ initialization operation places on the system. If you define these logical names in the system logical name table, they are translated each time an AIJ file is initialized. The RDM$BIND_AIJ_INITIALIZE_IO_COUNT logical name specifies the number of asynchronous I/O operations that are queued at once to the AIJ file. If the logical name is not defined, the default value is 15, the minimum value is 1, and the maximum value is 32. The RDM$BIND_AIJ_INITIALIZE_IO_SIZE logical name controls the number of 512-byte disk blocks to be written per I/O operation. If the logical name is not defined, the default value is 127, the minimum value is 4, and the maximum value is 127. Reducing the value of either logical will probably increase the amount of time needed to initialize the AIJ file after a backup. However, it may also reduce load on the I/O subsystem. o You should use the density values added in OpenVMS Version 7.2-1 for OpenVMS tape device drivers that accept them because previously supported values may not work as expected. If previously supported values are specified for drivers that support the OpenVMS Version 7.2-1 density values, the older values are translated to the Version 7.2-1 density values if possible. If the value cannot be translated, a warning message is generated, and the specified value is used. If you use density values added in OpenVMS Version 7.2-1 for tape device drivers that do not support them, the values are translated to acceptable values if possible. If the value cannot be translated, a warning message is generated and the density value is translated to the existing default internal density value (MT$K_DEFAULT). One of the following density-related errors is generated if there is a mismatch between the specified density value and the values that the tape device driver accepts: %DBO-E-DENSITY, TAPE_DEVICE:[000000]DATABASE.BCK; does not support specified density %DBO-E-POSITERR, error positioning TAPE_DEVICE: %DBO-E-BADDENSITY, The specified tape density is invalid for this device o If you want to use an unsupported density value, use the VMS INITIALIZE and MOUNT commands to set the tape density. Do not use the Density qualifier. o When you use the RMU Backup After_Journal command with the Log qualifier, the DCL global symbol RDM$AIJ_LAST_OUTPUT_FILE is automatically created. The value of the symbol is the full output backup AIJ file specification. o Because data stream names representing the database are generated based on the backup file name specified for the Oracle RMU backup command, you must either use a different backup file name to store the next backup of the database to the Librarian utility or first delete the existing data streams generated from the backup file name before the same backup file name can be reused. To delete the existing data streams stored in the Librarian utility, you can use a Librarian management utility or the Oracle RMU Librarian/Remove command. o The system logical RDM$BIND_AIJBCK_CHECKPOINT_TIMEOUT can be configured to control the checkpoint stall duration independent of the AIJ shutdown parameter. This logical works for both the AIJ backup and Automatic Backup Server (ABS) utilities. 4 Examples Example 1 Assuming that you have enabled after-image journaling for the MF_ PERSONNEL database, the following command causes extensible .aij entries to be backed up continuously until the time specified. $ RMU/BACKUP/AFTER_JOURNAL/CONTINUOUS/THRESHOLD=500 - _$ /INTERVAL=300/UNTIL="01-JUL-1996 16:15:00.00" - _$ MF_PERSONNEL.RDB DISK12:[PERS_AIJ]BU_PERSONNEL.AIJ Every 300 seconds, the backup process tests to determine if the active .aij file on disk has reached the threshold size of 500 blocks. If not, transaction processing continues normally for one or more 300-second intervals until the threshold test indicates that the active .aij file has reached a size of at least 500 blocks. When the .aij file reaches that file size, Oracle RMU allows existing transactions to continue to write to the active .aij file but does not allow new transactions to start. Assuming that the active .aij file contains 550 blocks, Oracle Rdb moves those 550 blocks to the backup journal file and deletes them from the active journal file. Then, the backup process determines if the transactions already in progress have written more journal records to the active journal file during the backup operation. If so, Oracle RMU moves those journal records to the backup file. After Oracle Rdb completely moves the active journal file, it truncates the journal file to 0 blocks. Oracle Rdb then allows new transactions to start and the backup process resumes threshold testing at 300-second intervals. The backup process continues until the time and date specified by the Until qualifier. Example 2 The following examples show backing up .aij files in sequence. Note that a number of transactions were committed to the database between backup operations. $ RMU/BACKUP/AFTER_JOURNAL/LOG MF_PERSONNEL MFPERS_BKUP_AIJ1.AIJ %RMU-I-AIJBCKBEG, beginning after-image journal backup operation %RMU-I-OPERNOTIFY, system operator notification: Oracle Rdb V7.2 Database DISK1:[DB]MF_PERSONNEL.RDB;1 Event Notification AIJ backup operation started %RMU-I-AIJBCKSEQ, backing up after-image journal sequence number 0 %RMU-I-LOGBCKAIJ, backing up after-image journal AIJ1 at 16:35:53.41 %RMU-I-LOGCREBCK, created backup file DISK1:[DB]MFPERS_BKUP_AIJ1.AIJ;1 %RMU-I-AIJBCKSEQ, backing up after-image journal sequence number 1 %RMU-I-LOGBCKAIJ, backing up after-image journal AIJ2 at 16:35:54.58 %RMU-I-QUIETPT, waiting for database quiet point %RMU-I-OPERNOTIFY, system operator notification: Oracle Rdb V7.2 Database DISK1:[DB]MF_PERSONNEL.RDB;1 Event Notification AIJ backup operation completed %RMU-I-AIJBCKEND, after-image journal backup operation completed successfully %RMU-I-LOGAIJJRN, backed up 2 after-image journals at 16:35:56.40 %RMU-I-LOGAIJBLK, backed up 508 after-image journal blocks at 16:35:56.41 . . . $ More transactions committed to the database . . . $ RMU/BACKUP/AFTER_JOURNAL/LOG MF_PERSONNEL MFPERS_BKUP_AIJ2.AIJ %RMU-I-AIJBCKBEG, beginning after-image journal backup operation %RMU-I-OPERNOTIFY, system operator notification: Oracle Rdb V7.2 Database DISK1:[DB]MF_PERSONNEL.RDB;1 Event Notification AIJ backup operation started %RMU-I-AIJBCKSEQ, backing up after-image journal sequence number 2 %RMU-I-LOGBCKAIJ, backing up after-image journal AIJ1 at 16:47:44.66 %RMU-I-LOGCREBCK, created backup file DISK2:[AIJ]MFPERS_BKUP_AIJ2.AIJ;1 %RMU-I-OPERNOTIFY, system operator notification: Oracle Rdb V7.2 Database DISK1:[DB]MF_PERSONNEL.RDB;1 Event Notification AIJ backup operation completed %RMU-I-AIJBCKEND, after-image journal backup operation completed successfully %RMU-I-LOGAIJJRN, backed up 1 after-image journal at 16:47:46.57 %RMU-I-LOGAIJBLK, backed up 254 after-image journal blocks at 16:47:46.57 Example 3 The following example uses the Edit_Filename qualifier to give the .aij backup file a meaningful file name. The Rename qualifier specifies that Oracle RMU should create the backup file by renaming the current .aij file and by creating a new .aij file with the same name as the original .aij file. $ RMU/BACKUP/AFTER_JOURNAL MF_PERSONNEL - _$ /EDIT_FILENAME=(SEQUENCE,"_",HOUR,"_",MINUTE,"_",MONTH,"_", - _$ DAY_OF_MONTH) AIJ2/RENAME $ DIR DISK1:[DB.AIJ2]*.AIJ Directory DISK1:[DB.AIJ_TWO] AIJ23_15_46_07_09.AIJ;1 Example 4 The following example shows the syntax to use when you want the .aij backup file name to default to that previously specified with the RMU Set After_Journal command. Note that the .aij backup file name used is that which corresponds to the first .aij file included in the backup operation. $ RMU/SET AFTER_JOURNAL MF_PERSONNEL /ENABLE/RESERVE=5 - _$ /ADD=(NAME=AIJ1, FILE=DISK1:[AIJ]AIJ_ONE, - _$ BACKUP_FILE=DISK4:[AIJBCK]AIJ1BCK) - _$ /ADD=(NAME=AIJ2, FILE=DISK2:[AIJ]AIJ_TWO, - _$ BACKUP_FILE=DISK4:[AIJBCK]AIJ2BCK) - _$ /ADD=(NAME=AIJ3, FILE=DISK3:[AIJ]AIJ_THREE, - _$ BACKUP_FILE=DISK4:[AIJBCK]AIJ3BCK) %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery $ ! $ !Assume backup operation was performed and other database activity occurs. $ !Then back up the .aij files: $ ! $ RMU/BACKUP/AFTER_JOURNAL MF_PERSONNEL.RDB "" $ ! $ DIR DISK4:[AIJBCK] Directory DISK4:[AIJBCK] AIJ1BCK.AIJ;1 Example 5 The following example uses a density value with compression: RMU/BACKUP/AFTER_JOURNAL /DENSITY=(TK89,COMPACTION)/REWIND - /LABEL=(LABEL1,LABEL2) MF_PERSONNEL TAPE1:MFP.AIJ, TAPE2: 3 Plan Executes a backup plan file previously created with the RMU Backup command (or created manually by the user). 4 Description A backup plan file is created when you execute an RMU Backup command with the Parallel and List_Plan qualifiers. See Backup Database for details on creating a plan file and the format of a plan file. 4 Format (B)0RMU/Backup/Plan plan-file Command Qualifiers x Defaults x /[No]Execute x Execute /List_Plan=output-file x None 4 Parameters 5 plan-file-spec The file specification for the backup plan file. The default file extension is .plan. 4 Command_Qualifiers 5 Execute Execute Noexecute The Execute qualifier specifies that Oracle RMU is to execute the plan file. The Noexecute qualifier specifies that Oracle RMU should not execute the plan file, but instead perform a validity check on the contents of the plan file. The validity check determines such things as whether the storage areas names assigned to each worker executor exist. By default, Oracle RMU executes the backup plan file when the RMU Backup Plan command is issued. 5 List_Plan List_Plan=output-file Specifies that Oracle RMU should generate a new plan file and write it to the specified output file. This new plan file is identical to the plan file you specified on the command line (the "original" plan file) with the following exceptions: o Any user-added comments in the original plan file do not appear in the new plan file. o The new plan file is formatted to match the standard format for RMU Backup plan files. 4 Usage_Notes o To use the RMU Backup Plan command for a database, you must have the RMU$BACKUP privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o To execute the RMU Backup Plan command, Oracle SQL/Services must be installed on your system. 4 Examples Example 1 The following example first creates a plan file by issuing an RMU Backup command with the Parallel and List_Plan qualifiers. Oracle RMU does not execute the plan file because the Noexecute qualifier is specified. The second command issues the RMU Backup Plan command to execute the plan file created with the RMU Backup command. $ ! Create the Backup plan file: $ ! $ RMU/BACKUP/PARALLEL=(EXEC=4, NODE=(NODE1, NODE2)) - _$ /LIST_PLAN=(PARTIAL.PLAN)/NOEXECUTE/INCLUDE=(RDB$SYSTEM, - _$ EMPIDS_LOW, EMPIDS_MID, EMPIDS_OVER, SALARY_HISTORY, EMP_INFO) - _$ /LABEL=(001, 002, 003, 004, 005, 006, 007, 008, 009) - _$ /CHECKSUM_VERIFICATION - _$ MF_PERSONNEL TAPE1:MF_PARTIAL.RBF, TAPE2:, TAPE3, TAPE4 $ ! $ ! Execute the plan file created with the previous command: $ ! $ RMU/BACKUP/PLAN partial.plan 2 Checkpoint When fast commit is enabled, requests that each active database process on each node flush updated database pages from its buffer pool to disk. 3 Description Usually, each process performs a checkpoint operation after a certain set of thresholds has been exceeded. The RMU Checkpoint command allows you to spontaneously force each process to perform a checkpoint operation. Performing a checkpoint operation is useful for several purposes. A checkpoint operation with the Wait qualifier causes all updated database pages to be flushed to disk. A checkpoint operation also improves the redo performance of the database recovery (DBR) process (although the per-process parameters should have already been properly initialized with this goal in mind). When the Checkpoint command with the Wait qualifier completes (the system prompt is returned), all active processes have successfully performed a checkpoint operation. When the system prompt is returned after you issue the Checkpoint command with the Nowait qualifier, there is no guarantee that all active processes have successfully performed a checkpoint operation. 3 Format (B)0RMU/Checkpoint root-file-spec Command Qualifiers x Default x /[No]Wait[/Until=date-and-time] x /Wait 3 Parameters 4 root-file-spec The root file specification for the database you want to checkpoint. You can use either a full or partial file specification, or a logical name. If you specify only a file name, Oracle Rdb looks for the database in the current default directory. If you do not specify a file extension, Oracle Rdb assumes a file extension of .rdb. 3 Command_Qualifiers 4 Wait Wait[/Until] Nowait Specifies whether or not the system prompt is to be returned before the checkpoint operation completes. When you specify the Wait qualifier without the Until qualifier, the system prompt is not returned to you until all processes have flushed updated database pages to disk. The Wait qualifier is the default. Used with the Wait qualifier, the Until qualifier specifies the time at which the RMU Checkpoint/Wait command stops waiting for the checkpoint and returns an error message. If you do not specify the Until qualifier, the wait is indefinite. When you specify the Nowait qualifier, the system prompt is returned immediately, before all processes have flushed database pages to disk. In addition, when you specify the Nowait qualifier, there is no guarantee that all processes will flush their database pages to disk. The Nowait qualifier is useful when it is more essential that the system prompt be returned immediately than it is to be certain that all processes have checkpointed. 3 Usage_Notes o To use the RMU Checkpoint command for a database, you must have the RMU$BACKUP or RMU$OPEN privilege in the root file access control list (ACL) for the database or you must have the OpenVMS WORLD privilege. o The RMU Checkpoint command is useful only if the database fast commit feature has been enabled. If the fast commit feature is disabled, this command does nothing. For more information on the fast commit feature, see the FAST COMMIT IS ENABLED section of the SQL ALTER DATABASE statement in the Oracle Rdb SQL Reference Manual. 3 Examples Example 1 The following command causes all the active database processes on all nodes to immediately perform a checkpoint operation: $ RMU/CHECKPOINT MF_PERSONNEL.RDB Example 2 The following command requests that all the active database processes on all nodes perform a checkpoint operation and that the system prompt be returned to you immediately. In this case, there is no guarantee that all processes will actually perform a checkpoint operation. $ RMU/CHECKPOINT/NOWAIT MF_PERSONNEL.RDB 2 Close Closes an open database. You should always specify the Wait qualifier, unless you are attempting to recover from some failure. When you specify the Wait qualifier, Oracle RMU performs all the auxiliary actions required to close and recover the database clusterwide, and it does not return the system prompt until those actions have been completed. If you use the RMU Close command with the Nowait qualifier, the database must be open on the node where you issue the command. Otherwise, you will receive an error message stating that the database is not known. The system prompt is returned immediately, but it is only an indication that the database will be closed as soon as all other users have finished accessing the database. Therefore, the Wait qualifier is used almost exclusively. 3 Description The RMU Close command closes an open database. A database root file is considered open if it has been specified in a previous RMU Open command or has active users attached to it. You can close the database immediately by specifying the Abort qualifier, or you can allow current users to finish their session by specifying the Noabort qualifier. If you have specified manual opening for your database (with the OPEN IS MANUAL clause of the SQL ALTER DATABASE statement), you must use the RMU Open command to manually open the database before any users can invoke it and the RMU Close command to manually close the database. If you have specified automatic opening for your database (with the OPEN IS AUTOMATIC clause of the SQL ALTER DATABASE statement), the RMU Close command affects current database users only. Current processes are detached from the database but they and new processes can immediately reattach to the database. Use the RMU Show Users command to display information about databases currently in use on your node. Use the RMU Dump Users command to display information about databases currently in use on your cluster. 3 Format (B)0 RMU/Close root-file-spec [,...] Command Qualifiers x Defaults x /[No]Abort=option x /Abort=Forcex /[No]Cluster x See description /Path x None /[No]Statistics=Export x /Nostatistics /[No]Wait x /Nowait 3 Parameters 4 root-file-spec root-file-spec[,...] An open database root file. The default file extension is .rdb. 3 Command_Qualifiers 4 Abort Abort=option Noabort Specifies whether to close the database immediately or allow processes to complete. The Abort qualifier has two options. Both refer to OpenVMS system services. The options are as follows: o Forcex When you use the Forcex (forced exit) option, recovery units are recovered and no recovery-unit journal (.ruj) files are left in the directories. Therefore, the RMU Backup command works. The option cannot force an exit of a database process with a spawned subprocess or a suspended or swapped out process. It aborts batch jobs that are using the database. Forcex is the default. o Delprc When you use the Delprc (delete process) option, recovery units are not recovered. The .ruj files are left in the directories to be recovered on the next invocation of the database. The processes and any subprocesses of all database users are deleted, thereby deleting the processes from the database. Therefore, if you attempt to issue an RMU Backup command, you might receive the following error message: %RMU-F-MUSTRECDB, database must be closed or recovered The Delprc and Forcex options are based on OpenVMS system services $DELPRC and $FORCEX. Refer to the OpenVMS documentation set for more information. With the Noabort option, users already attached to the database can continue, and the root file global sections remain mapped in the virtual address file contents until all users exit the database. No new users will be allowed to attach to the database. When all current images terminate, Oracle RMU closes the database. 4 Cluster Cluster Nocluster Specifying the Cluster qualifier causes Oracle RMU to attempt to close a database on all nodes in a clustered environment that currently have the database open. Specifying the Cluster qualifier is similar to issuing the RMU Close command on every node in the cluster. Specifying the Nocluster qualifier causes Oracle RMU to close a database only on the cluster node from which you issue the RMU Close command. The default is the Cluster qualifier if you specify the Wait qualifier. The default is the Nocluster qualifier if you specify the Nowait qualifier. The following list describes the behavior of the command when you use various combinations of the [No]Cluster and [No]Wait qualifiers together in the same command line: o Cluster and Wait When you specify the Cluster and Wait qualifiers, the RMU Close command closes the database on every node in the cluster, even if the database is not opened on the node from which the command is issued. Because you specified the Cluster and Wait qualifiers, the RMU Close command closes and recovers the database on every node in the cluster before the DCL prompt is returned to you. o Cluster and Nowait When you specify the Cluster and Nowait qualifiers, the RMU Close command attempts to close the database on every node in the cluster. If the database is not opened on the node from which the Oracle RMU command is issued, the command cannot close the database on any node, and you receive the following error message: %RDMS-F-CANTCLOSEDB, database could not be closed as requested -RDMS-F-DBNOTACTIVE, database is not being used %RMU-W-FATALERR, fatal error on DISK1:[USER1]DATABASE.RDB;1 Because you used the Nowait qualifier, the database might not yet be closed on one or more nodes when the DCL prompt is returned to you. When you specify the Nowait qualifier, you can receive SYS-F-ACCONFLICT errors when you attempt to access a database after you have issued the RMU Close command with the Cluster and Nowait qualifiers and the DCL prompt has been returned, but the monitor has not yet closed the database on all nodes in the cluster. o Nocluster and Wait This combination provides the ability to have database shutdown complete on the local node before Oracle RMU returns to the DCL prompt. o Nocluster and Nowait When you specify the Nocluster and Nowait qualifiers, Oracle RMU closes the database only on the node from which you issue the command, regardless of whether or not the database is open on other nodes. Because you used the Nowait qualifier, the database might not yet be closed on the node from which you issued the command when the DCL prompt is returned to you. With the Nowait qualifier, you can receive SYS-F-ACCONFLICT errors when you attempt to access a database after you have issued the RMU Close command with the Cluster and Nowait qualifiers and the DCL prompt has been returned, but the monitor has not yet closed the database on all nodes in the cluster. 4 Path Specifies the full or relative data dictionary path name in which the definitions reside for the database you want to close. The Path qualifier is a positional qualifier. Positional qualifiers operate on specific parameters based on the placement of the qualifiers in the command line. The path name cannot include wildcard characters. 4 Statistics=Export Statistics=Export Nostatistics Specifies that statistic information is to be saved when the database is closed. The default is Nostatistics, which indicates that statistic information is not preserved when the database is closed. Clusterwide statistic information is not stored in the statistic file, which allows you to decide on which nodes the statistic information should be initially loaded when the database is opened. The statistic information is stored in a node-specific database file located in the same directory as the database root file. The file has the same name as the root-file-spec, with a default file extension of .rds. Because the statistic files contain node- specific information, they cannot be renamed or copied. They can be deleted if they are no longer needed. The Statistics=Export qualifier cannot be specified in conjunction with the Cluster qualifier. To preserve the statistics information for a database open on a cluster, you must specifically close the individual nodes. The RMU Backup command does not save the statistics files. They are considered temporary files and not part of the database. 4 Wait Wait Nowait Specify the Wait qualifier to cause Oracle RMU to close and recover the database before the system prompt is returned to you. The default is the Nowait qualifier. With the Nowait qualifier, the database might not be closed when the system prompt is returned to you. You can receive errors when you attempt to access a database after you issued the RMU Close command and the system prompt is returned, but before the monitor has closed the database. See the Usage Notes for restrictions on using the Wait qualifier. 3 Usage_Notes o To use the RMU Close command for a database, you must have the RMU$OPEN privilege in the root file access control list (ACL) for the database or the OpenVMS WORLD privilege. o To use the Wait qualifier, Oracle RMU requires that the database be recoverable for correct operation. It must be possible to attach to the database on a node where it is opened. There are database recovery failures that preclude further attaches to the database. When such a failure occurs, any attempt to attach to the database (for example, with an SQL ATTACH statement) causes the process to be deleted from the system. In other words, you are logged out. In this situation, the RMU Close command with the Wait qualifier has the same effect as the RMU Close command with the Cluster and Nowait qualifiers. The operation does not wait, and it does not close the database unless it is opened on the node from which you issue the Oracle RMU command. If you encounter this situation, enter the following command from a node on which the database is open to close the database: $ RMU/CLOSE/CLUSTER/NOWAIT/ABORT=DELPRC 3 Examples Example 1 When you issue the following command from a node in a cluster, the Cluster qualifier shuts down the database for the entire cluster, even if no users are on the node from which you issued the command. The Wait qualifier causes Oracle RMU to close the database before the system prompt is returned. $ RMU/CLOSE/CLUSTER/WAIT MF_PERSONNEL.RDB Example 2 The following command closes the mf_personnel database in the [.WORK] directory, all the databases in the [.TEST] directory, and the databases specified by the path names CDD$TOP.FINANCE and SAMPLE_DB: $ RMU/CLOSE DISK1:[WORK]MF_PERSONNEL, CDD$TOP.FINANCE/PATH, - _$ DISK1:[TEST]*, SAMPLE_DB/PATH 2 Collect_Optimizer_Statistics Collects cardinality and storage statistics for the Oracle Rdb optimizer. Also collects workload statistics if a workload profile has been generated. 3 Description The purpose of collecting optimizer statistics is to maintain up-to-date statistics that the Oracle Rdb optimizer uses to determine solution costs and cardinalities during query optimization. You can collect cardinality and storage statistics by issuing the RMU Collect Optimizer_Statistics command. You can direct Oracle RMU to collect these statistics for particular tables or indexes by using the Tables, System_Relations, or Indexes qualifiers. Before you can collect workload statistics, you must first generate a workload profile with SQL. The following list describes the general procedure for generating a workload profile and collecting workload statistics: 1. Enable workload profiling with the WORKLOAD COLLECTION IS ENABLED clause of the SQL ALTER DATABASE or SQL CREATE DATABASE statement. SQL creates a new system table called RDB$WORKLOAD. 2. Execute the queries for which you want the Oracle Rdb optimizer to have the best possible statistics. When you execute the queries, the optimizer determines which groups of columns are important for optimal processing of the query. These groups of columns are referred to as workload column groups. Note that a workload column group may actually contain only one column. Each set of workload column groups is entered as a row in the RDB$WORKLOAD system table. At this point, the only data in the RDB$WORKLOAD system table are the workload column groups, the tables with which the column group is associated, and the date they were entered into the table. No statistics are currently recorded in the RDB$WORKLOAD system table. 3. In most cases, now you disable workload profiling with the SQL ALTER DATABASE WORKLOAD COLLECTION IS DISABLED clause. Queries executed after you disable workload profiling are not scanned by the Oracle Rdb optimizer for workload column groups. You can leave the workload profiling enabled if the same queries are always executed. In such a case, no new rows are entered into the RDB$WORKLOAD system table. However, if you anticipate that queries will be executed for which you do not want workload profiling to be enabled, you need to disable workload collection. 4. Execute an RMU Collect Optimizer_Statistics command with the Statistics=(Workload) qualifier. Oracle RMU reads the RDB$WORKLOAD system table to determine for which column groups it should collect statistics, and then collects the statistics. 5. Execute the queries previously profiled again. The optimizer uses the statistics gathered by Oracle RMU to make a best effort at optimizing the profiled queries. The following list provides some guidelines on when to issue the RMU Collect Optimizer_Statistics command and which Statistics qualifier options you should use: o You should enable workload profiling and execute the RMU Collect Optimizer_Statistics command with the Statistics=(Workload) qualifier when you introduce new, complex, frequently used queries as part of your regular work. o You should execute the RMU Collect Optimizer_Statistics command with the Statistics=(Storage) qualifier after you add metadata, such as new tables or indexes, to the database. In this case, you do not need to reenable workload profiling. o You should execute the RMU Collect Optimizer_Statistics command with the Statistics=(Storage, Workload) qualifier when the data in the database has significantly increased, decreased, or changed. In this case, you do not need to reenable workload profiling. The statistics you can gather with the RMU Collect Optimizer_ Statistics command and a description of how the optimizer uses these statistics are summarized in Statistics Gathered by the RMU Collect Optimizer_Statistics Command. Table 6 Statistics Gathered by the RMU Collect Optimizer_ Statistics Command Cardinality Statistics Statistic Gathered: Definition: Used by Optimizer to: Table Number of rows in Determine solution cardinality. Cardinality a table. Index Number of Estimate the number of index Cardinality distinct key keys returned. values in an index. Index Prefix Number of Estimate the number of index Cardinality distinct key keys returned based on a sorted values in leading index range. parts of a multi- segmented B-tree index. Workload Statistics Statistic Gathered: Definition: Used by Optimizer to: Column Group Average number Determine strategies for Duplicity of duplicates equiselections (selections Factor per distinct with the IS NULL predicate value in a column or selection predicates with group. This is an the equals (=) operator), estimated value. equijoins, grouped aggregation (for example, the SQL GROUP BY clause), or projection operations (for example, the SQL DISTINCT clause). Column Group Number of table Estimate the effects of Null Factor rows with a NULL NULL data on equijoins and value in at least equiselections (because they one column of a imply the removal of rows with column group. NULL values). Also used for This is an estimating the cardinality of estimated value. an outer join result. Storage Statistics Statistic Gathered: Definition: Used by Optimizer to: Average Average number Estimate the cost of descending Index Depth of levels to the B-tree. (A cross join with (sorted traverse on a an inner table that is accessed indexes B-tree descent. by a sorted index involves only) repetitive B-tree descents.) Index Key Average number of Improve the cost estimate Clustering I/Os required of performing an index-only Factor to read one retrieval for hashed and sorted index key and indexes. all associated dbkeys during a hashed key lookup or a B- tree index scan, excluding the B-tree descent. Index Data Average number Estimate the cost for fetching Clustering of I/Os required data rows from a sorted index Factor to fetch data scan or from a hash bucket. rows using dbkeys associated with an index key. Table Row The average Estimate the cost of performing Clustering number of I/Os a sequential scan of a table. Factor required to read one row during a sequential scan of a table. 3 Format (B)0RMU/Collect Optimizer_Statistics root-file Command Qualifiers x Defaults x /Exclude_Tables=(table-list) x None /[No]Indexes[=(index-list)] x /Indexes /[No]Log[=file-name] x Current DCL verify value /Row_Count=n x /Row_Count=100 /Statistics[=(options)] x /Statistics /[No]System_Relations x /Nosystem_Relations /[No]Tables[=(table-list)] x /Tables /Transaction_Type=option x /Transaction_Type=Automatic 3 Parameters 4 root-file-spec root-file-spec Specifies the database for which statistics are to be collected. The default file type is .rdb. 3 Command_Qualifiers 4 Exclude_Tables Exclude_Tables Exclude_Tables=(table-list) Specifies a list of database tables to be excluded from statistics collection and update for statistics used by the Rdb query optimizer. You must specify at least one list. You can specify an options file in place of a list of tables. If the Exclude_Tables qualifier is used with the Tables qualifier in the same RMU Collect Optimizer command, the Exclude_Tables qualifier takes precedence. If the same table is specified in the table list for both qualifiers, that table is excluded from the statistics collection and update. 4 Indexes Indexes Indexes[=(index-list)] Noindex Specifies the index or indexes for which statistics are to be collected. If you do not specify an index-list, statistics for all indexes defined for the tables specified with the Tables qualifier are collected. If you specify an index-list, statistics are collected only for the named indexes. If you specify the Noindex qualifier, statistics for the index cardinality, average index depth, index key clustering factor, and index data clustering factor are not collected. Specify the Notable qualifier if you do not want statistics collected for tables. (Remember, the Tables qualifier without a table-list is the default.) The default is the Indexes qualifier without an index-list. 4 Log Log Log=file-name Nolog Specifies how the values calculated for the statistics are to be logged. Specify the Log qualifier to have the information displayed to SYS$OUTPUT. Specify the Log=file-spec qualifier to have the information written to a file. Specify the Nolog qualifier to prevent display of the information. If you do not specify any of variation of the Log qualifier, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 4 Row_Count Row_Count=n Specifies the number of rows that are sent in a single I/O request when Workload Statistics are collected. You can experiment to find the value for n that provides the best performance and memory usage for your database and environment. As you increase the value of n, you see an increase in performance at the expense of additional memory and overhead. The minimum value you can specify for n is 1. The default value for n is 100. 4 Statistics Statistics Statistics[=(options)] Specifies the type of statistics you want to collect for the items specified with the Tables, System_Relations, and Indexes qualifiers. If you specify the Statistics qualifier without an options list, all statistics are collected for the items specified. If you specify the Statistics qualifier with an options list, Oracle RMU collects types of statistics described in the following list. If you specify more than one option, separate the options with commas and enclose the options within parenthesis. The Statistics qualifier options are: o Cardinality Collects the table cardinality for the tables specified with the Tables and System_Relations qualifiers and the index and index prefix cardinalities for the indexes specified with the Indexes qualifier. Because cardinalities are automatically maintained by Oracle Rdb, it is usually not necessary to collect cardinality statistics using the RMU Collect Optimizer_Statistics command unless you have previously explicitly disabled cardinality updates. o Workload Collects the Column Group, Duplicity Factor, and Null Factor workload statistics for the tables specified with the Tables and System_Relations qualifiers. o Storage Collects the following statistics: - Table Row Clustering Factor for the tables specified with the Tables qualifier - Index Key Clustering Factor, the Index Data Clustering Factor, and the Average Index Depth for the indexes specified with the Indexes qualifier See System Tables Used to Store Optimizer Statistics in the Usage_Notes entry for this command for information on the columns and tables used in the system relations to store these statistics. 4 System_Relations System_Relations Nosystem_Relations Specifies that optimizer statistics are to be collected for system tables (relations) and their associated indexes. If you do not specify the System_Relations qualifier, or if you specify the Nosystem_Relations qualifier, optimizer statistics are not collected for system tables or their associated indexes. Specify the Noindex qualifier if you do not want statistics collected for indexes defined on the system tables. The default is the Nosystem_Relations qualifier. 4 Tables Tables Tables[=(table-list)] Notables Specifies the table or tables for which statistics are to be collected. If you specify a table-list, statistics for those tables and their associated indexes are collected. If you do not specify a table-list, statistics for all tables and their associated indexes in the database are collected. If you do not specify the Table qualifier, statistics for all tables are collected. If you specify the Notables qualifier, statistics for for the table cardinality, table row clustering factor, column group duplicity factor, and column group null factor are not collected. Specify the Noindex qualifier if you do not want statistics collected for indexes. The Tables qualifier without a table-list is the default. 4 Transaction_Type Transaction_Type=option Allows you to specify the transaction mode for the transactions used to collect statistics. Valid options are: o Automatic o Read_Only o Noread_Only You must specify an option if you use this qualifier. If you do not use any form of this qualifier, the Transaction_ Type=Automatic qualifier is the default. This qualifier specifies that Oracle RMU is to determine the transaction mode used to collect statistics. If any storage area in the database (including those not accessed for collecting statistics) has snapshots disabled, the transactions used to collect data are set to read/write mode. Otherwise, the transactions to collect data are set to read-only mode. The Transaction_Type=Read_Only qualifier specifies the transactions used to collect statistics be set to read-only mode. When you explicitly set the transaction type to read- only, snapshots need not be enabled for all storage areas in the database, but must be enabled for those storage areas from which statistics are collected. Otherwise, you receive an error and the collect optimizer statistics operation fails. You might select this option if not all storage areas have snapshots enabled and you are collecting statistics on objects that are stored only in storage areas with snapshots enabled. In this case, using the Transaction_Type=Read_Only qualifier allows you to collect statistics and impose minimal locking on other users of the database. The Transaction_Type=Noread_Only qualifier specifies that the transactions used to collect statistics be set to read/write mode. You might select this option if you want to eradicate the growth of snapshot files that occurs during a read-only transaction and are willing to incur the cost of increased locking that occurs during a read/write transaction. 3 Usage_Notes o To use the RMU Collect Optimizer_Statistics command for a database, you must have the RMU$ANALYZE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o When you use the SQL ALTER DATABASE statement to set the RDB$SYSTEM storage area to read-only access for your database, the Oracle Rdb system tables in the RDB$SYSTEM storage area are also set to read-only access. When the Oracle Rdb system tables are set to read-only access: o Automatic updates to table and index cardinality are disabled. o Manual changes made to the cardinalities to influence the optimizer are not allowed. o The I/O associated with the cardinality update is eliminated. o For indexes, the cardinality value is the number of unique entries for an index that allows duplicates. If the index is unique, Oracle Rdb stores zero for the cardinality, and uses the table cardinality instead. For tables, the cardinality value is the number of rows in the table. Oracle Rdb uses the cardinality values of indexes and tables to influence decisions made by the optimizer. If the actual cardinality values of tables and indexes are different from the stored cardinality values, the optimizer's performance can be adversely affected. o As Oracle RMU performs the collect operation, it displays the maximum memory required to perform the operation. If the maximum amount required is not available, Oracle RMU makes adjustments to try to make use of the memory that is available. However, if after making these adjustments, memory is still insufficient, the collect operation skips the updates for the table causing the problem and continues with the operation. The skipped table is noted in the log file with the message, "Unable to allocate memory for ; default statistics values used." To avoid this problem, use the OpenVMS System Generation Utility (SYSGEN) to increase the VIRTUALPAGECNT parameter. o If you prefer not to update optimizer statistics all at once, you can divide the work into separate commands. Oracle Corporation recommends that you collect Cardinality and Storage statistics in one RMU Collect Optimizer_Statistics command; and collect Workload statistics in a second command. o You must decide if the improved performance provided by enabling and maintaining the workload profile is worth the cost. Generally speaking, it is worth the cost of maintaining this table for a stable set of queries that are run on a regular basis; it is not worth the cost of maintaining this table when the majority of your queries are ad hoc queries, each of which uses different access strategies. For example, if the majority of queries that access the EMPLOYEES table use the EMPLOYEE_ID as the selection criteria and the queries are using the same access strategy, you might want to maintain a workload profile for the EMPLOYEES table. However, if some queries access the EMPLOYEES table through the EMPLOYEE_ID, some through the LAST_NAME, and others through the STATE, in an unpredictable manner, the queries are using different access strategies for which you probably do not want to maintain a workload profile. o Index prefix cardinalities are cumulative values. For example, suppose an index contains three segments and the first segment has a cardinality of A; the second has a cardinality of B; and the third has a cardinality of C. Then the index prefix cardinality for the first segment is A; the index prefix cardinality for the second segment is A concatenated with B (A|B); and the index prefix cardinality for the third segment is A concatenated with B concatenated with C (A|B|C). Therefore, the prefix cardinality for last segment in an index is always equal to the total cardinality for the index. Likewise, if the index only contains one segment, the index prefix cardinality is equal to the total cardinality for the index. In these cases, because the index prefix cardinality is the same as the total index cardinality, Oracle RMU does not calculate an index prefix cardinality. Instead, Oracle RMU stores a value of "0" for the index prefix cardinality and the optimizer uses the value stored for the total index cardinality. o Cardinality statistics are automatically maintained by Oracle Rdb. Physical storage and Workload statistics are only collected when you issue an RMU Collect Optimizer_Statistics command. To get information about the usage of Physical storage and Workload statistics for a given query, define the RDMS$DEBUG_FLAGS logical name to be "O". For example: $ DEFINE RDMS$DEBUG_FLAGS "O" When you execute a query, if workload and physical statistics have been used in optimizing the query, you see a line such as the following in the command output: ~O: Workload and Physical statistics used o Detected asynchronous prefetch should be enabled to achieve the best performance of this command. Beginning with Oracle Rdb V7.0, by default, detected asynchronous prefetch is enabled for databases created under Oracle Rdb V7.0 or converted to V7.0. You can determine the setting for your database by issuing the RMU Dump command with the Header qualifier. If detected asynchronous prefetch is disabled, and you do not want to enable it for the database, you can enable it for your Oracle RMU operations by defining the following logicals at the process level: $ DEFINE RDM$BIND_DAPF_ENABLED 1 $ DEFINE RDM$BIND_DAPF_DEPTH_BUF_CNT P1 P1 is a value between 10 and 20 percent of the user buffer count. o You can delete entries from the workload profile with the RMU Delete Optimizer_Statistics command. See Delete_Optimizer_ Statistics for details. o You can display entries from the workload profile with the RMU Show Optimizer_Statistics command. See Show Optimizer_ Statistics for details. o System Tables Used to Store Optimizer Statistics provides a summary of the system tables in which statistics gathered by the RMU Collect Optimizer_Statistics command are stored. Table 7 System Tables Used to Store Optimizer Statistics Statistic System Table Name Column Name Table RDB$RELATIONS RDB$CARDINALITY Cardinality Table Row RDB$RELATIONS RDB$ROW_CLUSTER_FACTOR Clustering Factor Column Group RDB$WORKLOAD RDB$DUPLICITY_FACTOR Duplicity Factor Column Group RDB$WORKLOAD RDB$NULL_FACTOR Null Factor Index RDB$INDICES RDB$CARDINALITY Cardinality Index Prefix RDB$INDEX_ RDB$CARDINALITY Cardinality SEGMENTS Average RDB$INDICES RDB$INDEX_DEPTH Index Depth (B-Trees only) Index Key RDB$INDICES RDB$KEY_CLUSTER_FACTOR Clustering Factor Index Data RDB$INDICES RDB$DATA_CLUSTER_FACTOR Clustering Factor 3 Examples Example 1 The following example collects cardinality statistics for the EMPLOYEES and JOB_HISTORY tables and their associated indexes. See the Usage Notes for an explanation for the value "0" for the index prefix cardinality. $ RMU/COLLECT OPTIMIZER_STATISTICS mf_personnel.rdb - _$ /STATISTICS=(CARDINALITY)/TABLES=(EMPLOYEES, JOB_HISTORY) - _$ /INDEXES=(EMP_LAST_NAME,EMP_EMPLOYEE_ID, EMPLOYEES_HASH, - _$ JH_EMPLOYEE_ID, JOB_HISTORY_HASH)/LOG Start loading tables... at 3-JUL-1996 09:35:25.19 Done loading tables.... at 3-JUL-1996 09:35:25.91 Start loading indexes... at 3-JUL-1996 09:35:25.92 Done loading indexes.... at 3-JUL-1996 09:35:26.49 Start collecting btree index stats... at 3-JUL-1996 09:35:28.17 Done collecting btree index stats.... at 3-JUL-1996 09:35:28.23 Start collecting table & hash index stats... at 3-JUL-1996 09:35:28.23 Done collecting table & hash index stats.... at 3-JUL-1996 09:35:28.52 Start calculating stats... at 3-JUL-1996 09:35:28.76 Done calculating stats.... at 3-JUL-1996 09:35:28.76 Start writing stats... at 3-JUL-1996 09:35:30.16 ---------------------------------------------------------------------- Optimizer Statistics collected for table : EMPLOYEES Cardinality : 100 Index name : EMP_LAST_NAME Index Cardinality : 83 Segment Column Prefix cardinality LAST_NAME 0 Index name : EMP_EMPLOYEE_ID Index Cardinality : 100 Segment Column Prefix cardinality EMPLOYEE_ID 0 Index name : EMPLOYEES_HASH Index Cardinality : 100 ---------------------------------------------------------------------- Optimizer Statistics collected for table : JOB_HISTORY Cardinality : 274 Index name : JH_EMPLOYEE_ID Index Cardinality : 100 Segment Column Prefix cardinality EMPLOYEE_ID 0 Index name : JOB_HISTORY_HASH Index Cardinality : 100 Done writing stats.... at 3-JUL-1996 09:35:30.83 Example 2 The following example collects storage statistics for the EMPLOYEES and JOB_HISTORY TABLES and their associated indexes: $ RMU/COLLECT OPTIMIZER_STATISTICS mf_personnel - _$ /STATISTICS=(STORAGE)/TABLES=(EMPLOYEES, JOB_HISTORY) - _$ /INDEXES=(EMP_LAST_NAME,EMP_EMPLOYEE_ID, EMPLOYEES_HASH, - _$ JH_EMPLOYEE_ID, JOB_HISTORY_HASH)/LOG Start loading tables... at 3-JUL-1996 10:28:49.39 Done loading tables.... at 3-JUL-1996 10:28:50.30 Start loading indexes... at 3-JUL-1996 10:28:50.30 Done loading indexes.... at 3-JUL-1996 10:28:51.03 Start collecting btree index stats... at 3-JUL-1996 10:28:53.27 Done collecting btree index stats.... at 3-JUL-1996 10:28:53.37 Start collecting table & hash index stats... at 3-JUL-1996 10:28:53.38 Done collecting table & hash index stats.... at 3-JUL-1996 10:28:53.80 Start calculating stats... at 3-JUL-1996 10:28:54.07 Done calculating stats.... at 3-JUL-1996 10:28:54.07 Start writing stats... at 3-JUL-1996 10:28:55.61 ---------------------------------------------------------------------- Optimizer Statistics collected for table : EMPLOYEES Row clustering factor : 0.2550000 Index name : EMP_LAST_NAME Average Depth : 2.0000000 Key clustering factor : 0.0481928 Data clustering factor : 1.1686747 Index name : EMP_EMPLOYEE_ID Average Depth : 2.0000000 Key clustering factor : 0.0100000 Data clustering factor : 0.9500000 Index name : EMPLOYEES_HASH Key clustering factor : 1.0000000 Data clustering factor : 1.0000000 -------------------------------------------------------------------- Optimizer Statistics collected for table : JOB_HISTORY Row clustering factor : 0.0930657 Index name : JH_EMPLOYEE_ID Average Depth : 2.0000000 Key clustering factor : 0.0500000 Data clustering factor : 0.9500000 Index name : JOB_HISTORY_HASH Key clustering factor : 1.0000000 Data clustering factor : 1.0000000 Done writing stats.... at 3-JUL-1996 10:28:56.41 Example 3 The following example enables workload collection with an SQL ALTER DATABASE statement, executes frequently run queries to generate a workload profile, collects workload statistics for the EMPLOYEES and JOB_HISTORY tables (along with their associated indexes), and then displays the statistics gathered. The SQL natural left outer join causes the first and third workload column groups to be created. The SQL DISTINCT clause causes the second and fourth workload column groups to be created. $ ! Enable workload collection: $ SQL SQL> ALTER DATABASE FILENAME mf_personnel.rdb cont> WORKLOAD COLLECTION IS ENABLED; SQL> -- SQL> -- Execute frequently run SQL queries. SQL> -- SQL> ATTACH 'FILENAME mf_personnel.rdb'; SQL> SELECT DISTINCT * cont> FROM JOB_HISTORY NATURAL LEFT OUTER JOIN EMPLOYEES; . . . SQL> DISCONNECT DEFAULT; SQL> -- Disable workload collection: SQL> ALTER DATABASE FILENAME mf_personnel.rdb cont> WORKLOAD COLLECTION IS DISABLED; SQL> EXIT; $ $ ! Direct Oracle RMU to collect statistics for the EMPLOYEES and $ ! JOB_HISTORY tables. $ ! $ RMU/COLLECT OPTIMIZER_STATISTICS mf_personnel.rdb - _$ /TABLE=(EMPLOYEES, JOB_HISTORY)/STATISTICS=(WORKLOAD)/LOG Start loading tables... at 3-JUL-1996 10:40:00.22 Done loading tables.... at 3-JUL-1996 10:40:00.90 Start collecting workload stats... at 3-JUL-1996 10:40:03.43 Maximum memory required (bytes) = 6810 Done collecting workload stats.... at 3-JUL-1996 10:40:05.03 Start calculating stats... at 3-JUL-1996 10:40:05.32 Done calculating stats.... at 3-JUL-1996 10:40:05.32 Start writing stats... at 3-JUL-1996 10:40:06.91 ---------------------------------------------------------------------- Optimizer Statistics collected for table : EMPLOYEES Workload Column group : EMPLOYEE_ID Duplicity factor : 1.0000000 Null factor : 0.0000000 Workload Column group : LAST_NAME, FIRST_NAME, MIDDLE_INITIAL, ADDRESS_DATA_1, ADDRESS_DATA_2, CITY, STATE, POSTAL_CODE, SEX, BIRTHDAY, STATUS_CODE Duplicity factor : 1.5625000 Null factor : 0.3600000 ---------------------------------------------------------------------- Optimizer Statistics collected for table : JOB_HISTORY Workload Column group : EMPLOYEE_ID Duplicity factor : 2.7040000 Null factor : 0.0000000 Workload Column group : EMPLOYEE_ID, JOB_CODE, JOB_START, JOB_END, DEPARTMENT_CODE, SUPERVISOR_ID Duplicity factor : 1.5420582 Null factor : 0.3649635 Done writing stats.... at 3-JUL-1996 10:40:07.46 Example 4 The following example collects all statistics (cardinality, workload, and storage) for all tables and indexes in the database except system relations. Output is written to the file stats_ nosys.log. $ RMU/COLLECT OPTIMIZER_STATISTICS mf_personnel.rdb - _$ /LOG=stats_nosys.log Example 5 The following example collects all statistics (cardinality, workload, and storage) for all tables, indexes, and system relations. Output is written to the file stats_all.log. $ RMU/COLLECT OPTIMIZER_STATISTICS mf_personnel.rdb/SYSTEM_RELATIONS - _$ /LOG=stats_all.log Example 6 In the following example the Employees and Departments tables are excluded from statistics collection. $ RMU/COLLECT OPTIMIZER_STATISTICS MF_PERSONNEL /LOG - _$ /EXCLUDE_TABLES=(EMPLOYEES,DEPARTMENTS) 2 Convert Converts any of the following versions (or any of the mandatory updates to these versions) of Oracle Rdb databases to an Oracle Rdb release 7.2 database: o Version 7.0 o Version 7.1 See the Oracle Rdb Installation and Configuration Guide for the proper backup procedure prior to installing a new release of Oracle Rdb and converting databases. NOTE The following are important issues to consider when you convert a database: o A database must be backed up immediately following an Oracle RMU convert operation. A database converted using the RMU Convert command may not be recoverable if a full database backup is not made immediately after the convert operation completes. If you attempt to restore a database using a backup file created prior to the conversion, the database may be left in an unrecoverable state. o If after-image journaling is enabled when you issue the Convert command, Oracle RMU disables after-image journaling during the convert operation and then does one of the following, depending on the type of .aij file or files being employed when the Convert command was issued: - If an extensible .aij file was being used, Oracle RMU creates a new journal for the converted database and enables after-image journaling. - If fixed-size .aij files were being used, Oracle RMU activates the next available fixed-size journal and enables after-image journaling. If another fixed- size journal is not available, journaling remains disabled. Use only the .aij file (or files) created or activated during or after the convert operation together with the backup file you created immediately after the convert operation to restore and recover your database. Any .aij files created prior to the Convert operation cannot be used to recover the converted database. If you issue an RMU Convert command with the Rollback qualifier, Oracle RMU disables after-image journaling and returns the message: RMU-I-CANTENAAIJ. Oracle Corporation recommends that you back up the database and enable after-image journaling when the convert operation completes. o Growth of the RDB$SYSTEM storage area is normal during a convert operation. You must be sure that there is sufficient disk space for the new metadata and the converted metadata. During a convert operation Oracle RMU makes an upgraded copy of the metadata. If the convert operation fails, the old metadata is available for rolling back. If you specify the Nocommit qualifier, both copies of the metadata exist at the same time (to allow a manual rollback operation). If you specify the Commit qualifier, the old metadata is deleted once the convert operation completes successfully. Read the Description help entry under this command carefully for important information on converting single-file and multifile databases. 3 Description The RMU Convert command operates by creating a converted copy of the system tables and indexes. This implies that the RDB$SYSTEM storage area might grow during the conversion, but it is unlikely that the system tables will be fragmented by the conversion process. Because a copy of the system tables is made, the time taken by the conversion is proportional to the amount of storage allocated to the system tables, or the number of rows in system tables, or both. This is typically a few minutes per database. However, if the database has very large system tables, the conversion can be costly. If the database has a large number of versions of some tables, it might be more efficient for you to use the SQL EXPORT and IMPORT statements to convert the database. After the conversion, both copies of the system tables are stored in the database. The Commit qualifier selects the converted copy and deletes the original copy. The Rollback qualifier selects the original copy and deletes the converted copy. You can specify either the Commit or the Rollback qualifier at a later time if you selected the Nocommit qualifier when the database was converted. Be aware that as long as Commit or Rollback are not selected after a Nocommit conversion, extra space will be taken up in the database to store both versions of the metadata. It is important to issue the Convert/Commit command after you have verified that the conversion was successful. (RMU will not let you convert to a newer version if the previous Convert was never committed, even if it was years ago.) While both copies of the system tables exist, the database is usable under Oracle Rdb release 7.2, but not under the earlier version. Also, DDL (data definition language) operations to the database are prohibited to ensure that both copies of the system tables remain consistent. After you specify either the Commit or the Rollback qualifier, you can again perform DDL operations on the database. If you convert a multifile database created prior to Oracle Rdb Version 6.1 by using the RMU Convert command with the Nocommit qualifier and then use the RMU Convert command with the Rollback qualifier to revert to the prior database structure level, subsequent verify operations might return an RMU-W-PAGTADINV warning message. See the Usage_Notes help entry under this command for details. 3 Format (B)0RMU/Convert database-list Command Qualifiers x Defaults x /[No]Commit x /Commit /[No]Confirm x See description /Path x None /Prefix_Collection=option x See description /Reserve = (Area=n, Aij=n) x See description /[No]Rollback x /Norollback 3 Parameters 4 database-list The database-list parameter is a list of databases to be converted. A list item can be either the file specification of a database root file or a data dictionary path name. You can use wildcards in the file specification of a database root file. You cannot use wildcards in a data dictionary path name. 3 Command_Qualifiers 4 Commit Commit Nocommit Makes the database conversion permanent. When you specify the Commit qualifier, the database is converted to an Oracle Rdb release 7.2 database and cannot be returned to the previous version. The default is Commit. When you specify the Nocommit qualifier, you can convert the database to Oracle Rdb release 7.2 and roll it back to the previous version at a later time. Using the Nocommit qualifier is helpful when you want to test your applications against a new version of Oracle Rdb. In the event that you find problems, you can roll back to the previous version. Once you feel confident that your applications work well with the new version, you should commit the converted database, otherwise unnecessary space is taken up in the database to store the obsolete alternate version of the metadata. 4 Confirm Confirm Noconfirm Requests user input during the conversion procedure. When you specify the Confirm qualifier, Oracle RMU asks if you are satisfied with your database and aij backup files. If the database being converted has after-image journaling enabled, Oracle RMU asks if you want to continue and states that after- image journaling will be temporarily disabled. 4 Path Path Identifies that the database is being specified by its data dictionary path name instead of its file specification. The Path qualifier is a positional qualifier. 4 Prefix_Collection Prefix_Collection=option When you convert a database to release 7.2 from a release of Oracle Rdb prior to release 7.0, you can use the Prefix_ Collection qualifier to specify that sorted index prefix cardinality collection be Enabled, Enabled Full, or Disabled for all sysem and user sorted indexes. The following options are available for use with the Prefix_ Collection qualifier: o Disabled Specifies that index prefix cardinality collection is to be disabled. o Enabled Specifies that default collection is performed. The Oracle Rdb optimizer collects approximate cardinality values for the index columns to help in future query optimization. Enabled Estimate Specifies that prefix cardinality values for all indexes are to be estimated. o Enabled Collect Specifies that prefix cardinality values for all indexes are to be collected by calling the RMU Collect command. o Full Requests that extra I/O be performed, if required, to ensure that the cardinality values reflect the key value changes of adjacent index nodes. o Full=Estimate Specifies that prefix cardinality values for all indexes are to be estimated. o Full=Collect Specifies that prefix cardinality values for all indexes are to be collected by calling the RMU Collect command. 4 Reserve Reserve=(Area=n,Aij=n) Reserves space in the database root file for storage areas or .aij files, or both. Replace the character n with the number of storage areas or .aij files for which you want to reserve space. Note that you cannot reserve areas for a single-file database. You can reserve .aij files for a single-file database, but once the database is converted, you cannot alter that reservation unless you backup and restore the database. This qualifier is useful if, when you are converting your database, you anticipate the need for additional storage areas or .aij files. Because the addition of new storage areas or .aij files requires that users not be attached to the database, adding them while the database is being converted minimizes the time that the database is inaccessible to users. By default, one .aij file and no storage area files are reserved. 4 Rollback Rollback Norollback Returns a database that has been converted to an Oracle Rdb release 7.2 database (but not committed) to the previous version. You might decide to return to the previous version of the database for technical, performance, or business reasons. The Norollback qualifier prevents you from returning your converted database to the previous version. The default is the Norollback qualifier. If you specify both the Nocommit qualifier and the Rollback qualifier in the same RMU Convert command, your database is converted to Oracle Rdb release 7.2 and immediately rolled back to the previous version when the RMU Convert command is executed. This qualifier is valid only if you are converting from one of the following releases: 7.0 or 7.1. 3 Usage_Notes o To use the RMU Convert command for a database, you must have the RMU$CONVERT or RMU$RESTORE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o The RMU Convert command requires read/write access to the database root file, the RDB$SYSTEM area, and the directory in which the .ruj file will be entered. o Oracle Corporation recommends that you update multisegment index cardinalities as part of, or soon after, the convert operation completes. Stored cardinality values can differ from the actual cardinality values if the RDB$SYSTEM storage area has been set to read-only access. If you use the Confirm and Commit qualifiers when you issue the RMU Convert command, Oracle RMU asks if you want to update multisegment index cardinalities with actual index values and provides an estimate on the time it will take to perform the update. If you choose not to update these cardinalities with actual values as part of the convert operation, or if you do not use the Confirm qualifier, Oracle RMU updates the multisegment index cardinalities with estimated values. In such a case, you should update the cardinalities with actual values as soon as possible by issuing an RMU Collect Optimizer_Statistics command. See Collect_Optimizer_Statistics for details. o If the database conversion does not complete (for example, because of a system failure or an Oracle Rdb monitor shutdown), you can execute the RMU Convert command again later. The ability to complete the conversion process later keeps you from having a half-converted database that is corrupted. o If the RDB$SYSTEM storage area attribute is set to read- only access, the RMU Convert command proceeds to reset the attribute to read/write, convert the database and then reset the attribute to read-only when the conversion is complete. If the RDB$SYSTEM storage area is located on a device that cannot be written to, the database conversion fails and returns an error message. o You are prompted to specify the Prefix_Collection parameters if the following conditions are true: o The Prefix_Collection qualifier is not specified. o The RMU Convert process is not running as a batch job. o The Noconfirm qualifier is not specified. As a response to the prompt, you can enter "E(NABLE)" for the equivalent of Prefix_Collection=Enabled, "F(ULL)" for the equivalent of Prefix_Collection=Full, "D(ISABLE)" for the equivalent of Prefix_Collection=Disabled, or the default of "I(GNORE)" if you do not want to change any prefix cardinality settings. 3 Examples Example 1 The first command in the following example converts an Oracle Rdb release 7.0 database with an extensible .aij file to an Oracle Rdb release 7.2 database. Because the Nocommit qualifier is specified in the first command, you can roll back the converted database (the Oracle Rdb release 7.2 database) to the original Oracle Rdb release 7.0 database. After-image journaling is disabled while the database is being converted. After the database is converted, a new extensible .aij file is created and after-image journaling is enabled again. Note that .aij files are version-specific. You should perform a full backup operation after a conversion because the old version and the new version of the .aij file are incompatible. In the second command, the converted database is rolled back to the original database. $RMU/CONVERT/CONFIRM/NOCOMMIT MF_PERSONNEL.RDB %RMU-I-RMUTXT_000, Executing RMU for Oracle Rdb V7.2-00 Are you satisfied with your backup of DISK1:[TESTS]MF_PERSONNEL.RDB;1 and your backup of any associated .aij files [N]? Y %RMU-I-AIJ_DISABLED, after-image journaling is being disabled temporarily for the Convert operation %RMU-I-LOGCONVRT, database root converted to current structure level %RMU-S-CVTDBSUC, database DISK1:[TESTS]MF_PERSONNEL.RDB;1 successfully converted from version V7.0 to V7.2 %RMU-I-LOGCREAIJ, created after-image journal file DISK1:[TESTS]BACKUP_AFTER1.AIJ;2 %RMU-I-LOGMODSTR, activated after-image journal "AFTER1" %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery $RMU/CONVERT/ROLLBACK MF_PERSONNEL.RDB %RMU-I-RMUTXT_000, Executing RMU for Oracle Rdb V7.2-00 Are you satisfied with your backup of DISK1:[TESTS]MF_PERSONNEL.RDB;1 and your backup of any associated .aij files [N]? Y %RMU-I-AIJ_DISABLED, after-image journaling is being disabled temporarily for the Convert operation %RMU-I-LOGCONVRT, database root converted to current structure level %RMU-I-CVTROLSUC, CONVERT rolled-back for DISK1:[TESTS]MF_PERSONNEL.RDB;1 to version V7.0 %RMU-I-CANTENAAIJ, The JOURNAL is now DISABLED. RMU CONVERT can not enable the JOURNAL for previous versions. You must do this manually. %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery Example 2 This example is the same as Example 1, except fixed-size .aij journals are being employed at the time of the conversion. The first command in this example converts an Oracle Rdb release 7.1 database with fixed-size .aij files to an Oracle Rdb release 7.2 database. Because the Nocommit qualifier is specified in the first command, you can roll back the converted database (the Oracle Rdb release 7.2 database) to the original Oracle Rdb V7.1 database. After-image journaling is disabled while the database is being converted. After the database is converted, Oracle RMU activates the next fixed-size .aij file and enables after-image journaling. Note that .aij files are version specific. You should perform a full backup operation after a conversion because the old .aij files are incompatible with the newly converted database. In the second command, the converted database is rolled back to the original database. $RMU/CONVERT/CONFIRM/NOCOMMIT MF_PERSONNEL.RDB %RMU-I-RMUTXT_000, Executing RMU for Oracle Rdb V7.2-00 Are you satisfied with your backup of DISK1:[TESTS]MF_PERSONNEL.RDB;1 and your backup of any associated .aij files [N]? Y %RMU-I-AIJ_DISABLED, after-image journaling is being disabled temporarily for the Convert operation %RMU-I-LOGCONVRT, database root converted to current structure level %RMU-S-CVTDBSUC, database DISK1:[TESTS]MF_PERSONNEL.RDB;1 successfully converted from version V7.1 to V7.2 %RMU-I-LOGMODSTR, activated after-image journal "AFTER2" %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery $RMU/CONVERT/ROLLBACK MF_PERSONNEL.RDB %RMU-I-RMUTXT_000, Executing RMU for Oracle Rdb V7.2-00 Are you satisfied with your backup of DISK1:[TESTS]MF_PERSONNEL.RDB;1 and your backup of any associated .aij files [N]? Y %RMU-I-AIJ_DISABLED, after-image journaling is being disabled temporarily for the Convert operation %RMU-I-LOGCONVRT, database root converted to current structure level %RMU-I-CVTROLSUC, CONVERT rolled-back for DISK1:[TESTS]MF_PERSONNEL.RDB;1 to version V7.1 %RMU-I-CANTENAAIJ, The JOURNAL is now DISABLED. RMU CONVERT can not enable the JOURNAL for previous versions. You must do this manually. %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery Example 3 The following command converts all the databases in DISK1:[RICK] and its subdirectories and also the SPECIAL_DB database that is identified by its data dictionary path name. The Noconfirm qualifier is specified, so Oracle RMU does not request user input. The Nocommit qualifier is not specified, so the default qualifier, Commit, is used by default and the converted databases cannot be rolled back. $ RMU/CONVERT/NOCONFIRM DISK1:[RICK...]*.RDB,CDD$TOP.RICK.SPECIAL_DB - _$ /PATH Example 4 The following command converts an Oracle Rdb release 7.0 database to release 7.2. In addition, it reserves space in the database root file of the converted database for four .aij files. After- image journaling is not enabled at the time the Convert command is issued. $RMU/CONVERT/CONFIRM/RESERVE=(AIJ=4)/COMMIT MF_PERSONNEL %RMU-I-RMUTXT_000, Executing RMU for Oracle Rdb V7.2-00 Are you satisfied with your backup of DISK1:[TESTS]MF_PERSONNEL.RDB;1 and your backup of any associated .aij files [N]? Y %RMU-I-LOGCONVRT, database root converted to current structure level %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery %RMU-S-CVTDBSUC, database DISK1:[TESTS]MF_PERSONNEL.RDB;1 successfully converted from version V7.0 to V7.2 Example 5 The following example shows how the contents of a batch file might look if you were to issue the RMU Convert command with the Confirm qualifier from a batch job. $ RMU/CONVERT/COMMIT/CONFIRM USER1:[COLLECT.V71DB]MF_PERSONNEL Y Y 2 Copy_Database Permits you to copy a database. 3 Description The RMU Copy_Database command allows you to modify certain area parameters when the copy operation is performed. All the files are processed simultaneously during the copy operation. The copy operation's performance is similar to that of the RMU Backup command. The RMU Copy_Database command eliminates the need for intermediate storage media. NOTE You must perform a full and complete Oracle RMU backup operation immediately after the Copy_Database operation completes to ensure that the database can be properly restored after a database failure or corruption. Also note that if you do not specify either the After_ Journal qualifier or the Aij_Options qualifier when you issue the RMU Copy_Database command, after-image journaling is disabled for the database copy and no .aij files are associated with the database copy. 3 Format (B)0RMU/Copy_Database root-file-spec [storage-area-list] Command Qualifiers x Defaults x /[No]After_Journal[=file-spec] x See description /[No]Aij_Options=journal-opts-file x See description /[No]Cdd_Integrate x Nocdd_Integrate /[No]Checksum_Verification x /Checksum_Verification /Close_Wait=n x See description /Directory=directory-spec x None /[No]Duplicate x /Noduplicate /Global_Buffers=global-buffer-options x Current value /Local_Buffers=local-buffer-options x Current value /Lock_Timeout=n x See description /[No]Log x Current DCL verify value /Nodes_Max=n x Current value /[No]Online x /Noonline /Open_Mode={Automatic|Manual} x Current value /Option=file-spec x None (B)0/Page_Buffers=n x n=33 /Path=cdd-path x Existing value /[No]Quiet_Point x /Quiet_Point /Root=file-spec x None /Transaction_Mode=(mode-list) x /Transaction_Mode=Current /Threads=n x /Threads=10 /Users_Max=n x Current value File or Area Qualifier x Defaults x /Blocks_Per_Page=n x None /Extension={Disable | Enable } x Current value /File=file-spec x None /Read_Only x Current value /Read_Write x Current value /Snapshots=(Allocation=n,File=file-spec) x None /[No]Spams x Current value /Thresholds=(n,n,n) x None 3 Parameters 4 root-file-spec The name of the database root file for the database you want to copy. 4 storage-area-list The name of one or more storage areas whose parameters you are changing. The storage-area-list parameter is optional. Unless you are using the RMU Copy_Database command to modify the parameters of one or more storage areas, you should not specify any storage area names. 3 Command_Qualifiers 4 After_Journal After_Journal[=file-spec] Noafter_Journal NOTE This qualifier is maintained for compatibility with versions of Oracle Rdb prior to Version 6.0. You might find it more useful to specify the Aij_Options qualifier, unless you are interested in creating an extensible .aij file only. Specifies how Oracle RMU is to handle after-image journaling and .aij file creation, using the following rules: o If you specify the After_Journal qualifier and provide a file specification, Oracle RMU enables journaling and creates a new extensible after-image journal (.aij) file for the database copy. o If you specify the After_Journal qualifier but you do not provide a file specification, Oracle RMU enables after-image journaling and creates a new extensible .aij file for the database copy with the same name as, but a different version number from, the .aij file for the database being copied. o If you specify the Noafter_Journal qualifier, Oracle RMU disables after-image journaling and does not create a new .aij file. o If you do not specify an After_Journal, Noafter_Journal, Aij_Options, or Noaij_Options qualifier, Oracle RMU disables after-image journaling and does not create a new .aij file. You can specify only one, or none, of the following after-image journal qualifiers in a single RMU Copy_Database command: After_ Journal, Noafter_Journal, Aij_Options, or Noaij_Options. You cannot use the After_Journal qualifier to create fixed-size .aij files; use the Aij_Options qualifier. 4 Aij_Options Aij_Options=journal-opts-file Noaij_Options Specifies how Oracle RMU is to handle after-image journaling and .aij file creation, using the following rules: o If you specify the Aij_Options qualifier and provide a journal-opts-file, Oracle RMU enables journaling and creates the .aij file or files you specify for the database copy. If only one .aij file is created for the database copy, it will be an extensible .aij file. If two or more .aij files are created for the database copy, they will be fixed-size .aij files (as long as at least two .aij files are always available). o If you specify the Aij_Options qualifier, but do not provide a journal-opts-file, Oracle RMU disables journaling and does not create any new .aij files. o If you specify the Noaij_Options qualifier, Oracle RMU disables journaling and does not create any new .aij files. o If you do not specify an After_Journal, Noafter_Journal, Aij_Options, or Noaij_Options qualifier, Oracle RMU disables after-image journaling and does not create a new .aij file. You can only specify one, or none, of the following after-image journal qualifiers in a single Oracle RMU command: After_Journal, Noafter_Journal, Aij_Options, Noaij_Options. See Show After_Journal for information on the format of a journal-opts-file. 4 Cdd_Integrate Cdd_Integrate Nocdd_Integrate Integrates the metadata from the root (.rdb) file of the database copy into the data dictionary (assuming the data dictionary is installed on your system). If you specify the Nocdd_Integrate qualifier, no integration occurs during the copy operation. You might want to delay integration of the database metadata with the data dictionary until after the copy operation finishes successfully. You can use the Nocdd_Integrate qualifier even if the DICTIONARY IS REQUIRED clause was used when the database being copied was defined. The Cdd_Integrate qualifier integrates definitions in one direction only-from the database file to the dictionary. The Cdd_Integrate qualifier does not integrate definitions from the dictionary to the database file. The Nocdd_Integrate qualifier is the default. 4 Checksum_Verification Checksum_Verification Nochecksum_Verification Requests that the page checksum be verified for each page copied. The default is to perform this verification. The Checksum_Verification qualifier uses significant CPU resources but can provide an extra measure of confidence in the quality of the data being copied. For offline copy operations, the additional CPU cost of using the Checksum_Verification qualifier might not be justified unless you are experiencing or have experienced disk, HSC, or CI port hardware problems. One symptom of these problems is pages being logged to the corrupt page table (CPT). For online copy operations, use of the Checksum_Verification qualifier offers an additional level of data security when the database employs disk striping or RAID (redundant arrays of inexpensive disks) technology. These technologies fragment data over several disk drives, and use of the Checksum_Verification qualifier permits Oracle RMU to detect the possibility that the data it is reading from these disks has been only partially updated. If you use either of these technologies, you should use the Checksum_Verification qualifier. Note, however, that if you specify the Nochecksum qualifier, and undetected corruptions exist in your database, the corruptions are included in the copied file. Such a corruption might be difficult to recover from, especially if it is not detected until weeks or months after the copy operation is performed. Overall, Oracle Corporation recommends that you use the Checksum_ Verification qualifier with all copy operations where integrity of the data is essential. 4 Close_Wait=n Specifies a wait time of n minutes before Oracle RMU automatically closes the database. You must supply a value for n. In order to use this qualifier, the Open_Mode qualifier on the RMU Copy_Database command line must be set to Automatic. 4 Directory Directory=directory-spec Specifies the default destination for the copied database files. Note that if you specify a file name or file extension, all copied files are given that file name or file extension. There is no default directory specification for this qualifier. See the Usage Notes for information on how this qualifier interacts with the Root, File, and Snapshot qualifiers and for warnings regarding copying database files into a directory owned by a resource identifier. If you do not specify this qualifier, Oracle RMU attempts to copy all the database files (unless they are qualified with the Root, File, or Snapshot qualifier) to their current location. 4 Duplicate Duplicate Noduplicate Causes the RMU Copy_Database command to generate a new database with the same content, but with a different identity from that of the original database. For this reason, .aij files cannot be interchanged between the original and the duplicate database. This qualifier creates copies of your databases that are expected to evolve independently in time. In this case, being able to exchange .aij files might be a security breach, and a likely source of corruption. A duplicate database has the same contents as the original database, but not the same identity. A database copied with the Noduplicate qualifier is an exact replica of the original database in every way and, therefore, .aij files can be interchanged between the original and duplicate database. The default is the Noduplicate qualifier. 4 Global_Buffers Global_Buffers=global-buffer-options Allows you to change the default global buffer parameters when you copy a database. The following options are available: o Disabled Use this option to disable global buffering for the copy of the original database. o Enabled Use this option to enable global buffering for the copy of the original database. You cannot specify both the Disabled and Enabled option in the same RMU Copy_Database command with the Global_Buffers qualifier. o Total=total-buffers Use this option to specify the number of buffers available for all users. o User_Limit=buffers-per-user Use this option to specify the maximum number of buffers available to each user. If you do not specify a global buffers option, the database is copied with the values that are in effect for the database you are copying. When you specify two or more options with the Global_Buffers qualifier, use a comma to separate each option and enclose the list of options in parentheses. 4 Local_Buffers Local_Buffers=local-buffer-options Allows you to change the default local buffer parameters when you copy a database. The following options are available: o Number=number-buffers Use this option to specify the number of local buffers that will be available for all users. You must specify a number between 2 and 32,767 for the number-buffers parameter. o Size=buffer-blocks Use this option to specify the size (specified in blocks) for each buffer. You must specify a number between 2 and 64 for the buffer-blocks parameter. If you specify a value smaller than the size of the largest page defined, Oracle RMU automatically adjusts the size of the buffer to hold the largest page defined. For example, if you specify the Local_Buffers=Size=8 qualifier and the largest page size for the storage areas in your database is 64 blocks, Oracle RMU automatically interprets the Local_Buffers=Size=8 qualifier as though it were a Local_Buffers=Size=64 qualifier. Take great care when selecting a buffer size; a poor choice causes performance to suffer greatly. The value specified for the buffer-blocks parameter determines the number of blocks for each buffer, regardless of whether local buffering or global buffering is enabled for the database. If you do not specify a Local_Buffers option, the database is copied with the values that are in effect for the database you are copying. 4 Lock_Timeout Lock_Timeout=n Specifies a timeout interval or maximum time in seconds to wait for the quiet-point lock and any other locks needed when the operation is performed online. When you specify the Lock_ Timeout=seconds qualifier, you must specify the number of seconds to wait for the quiet-point lock. If the time limit expires, an error is signaled and the copy operation fails. When the Lock_Timeout=seconds qualifier is not specified, the copy operation waits indefinitely for the quiet-point lock and any other locks needed during an online copy operation. The Lock_Timeout=seconds qualifier is ignored for offline copy operations. 4 Log Log Nolog Specifies whether the processing of the command is reported to SYS$OUTPUT. Specify the Log qualifier to request log output and the Nolog qualifier to prevent it. If you specify neither, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 4 Nodes_Max Nodes_Max=n Specifies a new value for the database maximum node count parameter for the database copy. The default is to leave the value unchanged. 4 Online Online Noonline Specifies that the copy database operation be performed while other users are attached to the database. The areas to be copied are locked for read-only access, so the operation is compatible with all but exclusive access. The default is the Noonline qualifier. 4 Open_Mode Open_Mode=Automatic Open_Mode=Manual Allows you to change the mode for opening a database when you copy a database. When you specify the Open_Mode=Automatic qualifier, users can invoke the database copy immediately after it is copied. If you specify the Open_Mode=Manual qualifier, an RMU Open command must be used to open the database before users can invoke the database copy. The Open_Mode qualifier also specifies the mode for closing a database. If you specify Open_Mode=Automatic, you can also use the Close_Wait qualifier to specify a time in minutes before the database is automatically closed. If you do not specify the Open_Mode qualifier, the database is copied with the open mode that is in effect for the database being copied. 4 Option Option=file-spec Specifies an options file containing storage area names, followed by the storage area qualifiers that you want applied to that storage area. Do not separate the storage area names with commas. Instead, put each storage area name on a separate line in the file. The storage area qualifiers that you can include in the options file are: Blocks_Per_Page, File, Snapshots, and Thresholds. You can use the DCL line continuation character, a hyphen (-), or the comment character (!) in the options file. There is no default for this qualifier. Example 6 in the Examples entry under this command shows an options file and how to specify it on the Oracle RMU command line. If the Option qualifier is specified, the storage-area-list parameter is ignored. 4 Page_Buffers Page_Buffers=n Specifies the number of buffers to be allocated for each database file to be copied. The number of buffers used is twice the number specified; half are used for reading the file and half for writing the copy. Values specified with the Page_Buffers qualifier can range from 1 to 5. The default value is 3. Larger values might improve performance, but they increase memory use. 4 Path Path=cdd-path Specifies a data dictionary path into which the definitions of the database copy will be integrated. If you do not specify the Path qualifier, Oracle RMU uses the CDD$DEFAULT logical name value of the user who enters the RMU Copy_Database command. If you specify a relative path name, Oracle Rdb appends the relative path name you enter to the CDD$DEFAULT value. If the cdd-path parameter contains nonalphanumeric characters, you must enclose it within quotation marks (""). Oracle Rdb ignores the Path qualifier if you use the Nocdd_ Integrate qualifier or if the data dictionary is not installed on your system. 4 Quiet_Point Quiet_Point Noquiet_Point Allows you to specify that a database copy operation is to occur either immediately or when a quiet point for database activity occurs. A quiet point is defined as a point where no active update transactions are in progress in the database. When you specify the Noquiet_Point qualifier, Oracle RMU proceeds with the copy operation as soon as the RMU Copy_Database command is issued, regardless of any update transaction activity in progress in the database. Because Oracle RMU must acquire concurrent-read locks on all physical and logical areas, the copy operation fails if there are any active transactions with exclusive locks on a storage area. However, once Oracle RMU has successfully acquired all concurrent-read storage area locks, it should not encounter any further lock conflicts. If a transaction that causes Oracle Rdb to request exclusive locks is started while the copy operation is proceeding, that transaction either waits or gets a lock conflict error, but the copy operation continues unaffected. If you intend to use the Noquiet_Point qualifier with a copy procedure that previously specified the Quiet_Point qualifier (or did not specify either the Quiet_Point or Noquiet_Point qualifier), you should examine any applications that execute concurrently with the copy operation. You might need to modify your applications or your copy procedure to handle the lock conflicts that can occur when you specify the Noquiet_Point qualifier. When you specify the Quiet_Point qualifier, the copy operation begins when a quiet point is reached. Other update transactions issued after the database copy operation begins are prevented from executing until after the root file for the database has been copied (copying of the database storage areas begins after the root file is copied). The default is the Quiet_Point qualifier. 4 Root Root=file-spec Requests that the database root file be copied to the specified location. See the Usage Notes for information on how this qualifier interacts with the Directory, File, and Snapshot qualifiers. 4 Transaction_Mode=(mode-list) Transaction_Mode=(mode-list) Sets the allowable transaction modes for the database root file created by the copy operation. The mode-list can include one or more of the following transaction modes: o All - Enables all transaction modes o Current - Enables all transaction modes that are set for the source database. This is the default transaction mode. o None - Disables all transaction modes o [No]Batch_Update o [No]Exclusive o [No]Exclusive_Read o [No]Exclusive_Write o [No]Protected o [No]Protected_Read o [No]Protected_Write o [No]Read_Only o [No]Read_Write o [No]Shared o [No]Shared_Read o [No]Shared_Write Your copy operation must include the database root file. Otherwise, RMU returns the CONFLSWIT error when you issue an RMU Copy_Database command with the Transaction_Mode qualifier. If you specify more than one transaction mode in the mode-list, enclose the list in parenthesis and separate the transaction modes from one another with a comma. Note the following: o When you specify a negated transaction mode such as Noexclusive_Write, it indicates that exclusive write is not an allowable access mode for the copied database. o If you specify the Shared, Exclusive, or Protected transaction mode, Oracle RMU assumes you are referring to both reading and writing in that transaction mode. o No mode is enabled unless you add that mode to the list, or you use the All option to enable all transaction modes. o You can list one transaction mode that enables or disables a particular mode followed by another that does the opposite. For example, Transaction_Mode=(Noshared_Write, Shared) is ambiguous because the first value disables Shared_Write access and the second value enables Shared_Write access. Oracle RMU resolves the ambiguity by first enabling the modes as specified in the modes-list and then disabling the modes as specified in the modes-list. The order of items in the list is irrelevant. In the example presented previously, Shared_Read is enabled and Shared_Write is disabled. 4 Threads=number Threads=number Specifies the number of reader threads to be used by the copy process. RMU creates so called internal 'threads' of execution to read data from one specific storage area. Threads run quasi-parallel within the process executing the RMU image. Each thread generates its own I/O load and consumes resources like virtual address space and process quotas (e.g. FILLM, BYTLM). The more threads, the more I/Os can be generated at one point in time and the more resources are needed to accomplish the same task. Performance increases with more threads due to parallel activities which keeps disk drives busier. However, at a certain number of threads, performance suffers because the disk I/O subsystem is saturated and I/O queues build up for the disk drives. Also the extra CPU time for additional thread scheduling overhead reduces the overall performance. Typically 2-5 threads per input disk drive are sufficient to drive the disk I/O susbsystem at its optimum. However, some controllers may be able to handle the I/O load of more threads, for example disk controllers with RAID sets and extra cache memory. In a copy operation, one thread moves the data of one storage area at a time. If there are more storage areas to be copied than there are threads, then the next idle thread takes on the next storage area. Storage areas are copied in order of the area size - largest areas first. This optimizes the overall elapsed time by allowing other threads to copy smaller areas while an earlier thread is still working on a large area. If no threads qualifier is specified, then 10 threads are created by default. The minimum is 1 thread and the maximum is the number of storage areas to be copied. If the user specifies a value larger than the number of storage areas, then RMU silently limits the number of threads to the number of storage areas. For a copy operation, you can specify a threads number as low as 1. Using a threads number of 1 generates the smallest system load in terms of working set usage and disk I/O load. Disk I/O subsystems most likely can handle higher I/O loads. Using a slightly larger value than 1 typically results in faster execution time. 4 Users_Max Users_Max=n Specifies a new value for the database maximum user count parameter. The default is to use the same value as is in effect for the database being copied. 4 Blocks_Per_Page Blocks_Per_Page=n Specifies a new page size for the storage area to which it is applied. You cannot decrease the page size of a storage area, and you cannot change the size of a storage area with a uniform page format. You might want to increase the page size in storage areas containing hash indexes that are close to full. By increasing the page size in such a situation, you prevent the storage area from extending. The Blocks_Per_Page qualifier is a positional qualifier. 4 Extension Extension=Disable Extension=Enable Allows you to change the automatic file extension attribute for a storage area when you copy a database. Use the Extension=Disable qualifier to disable automatic file extensions for a storage area. Use the Extension=Enable qualifier to enable automatic file extensions for a storage area. If you do not specify the Extension=Disable or the Extension=Enable qualifier, the storage areas are copied with the automatic file extension attributes that are in effect for the database being copied. The Extension qualifier is a positional qualifier. 4 File File=file-spec Requests that the storage area to which this qualifier is applied be copied to the specified location. See the Usage Notes for information on how this qualifier interacts with the Root, Snapshot, and Directory qualifiers and for warnings regarding copying database files into a directory owned by a resource identifier. The File qualifier is a positional qualifier. This qualifier is not valid for single-file databases. 4 Read_Only Use the Read_Only qualifier to change a read/write storage area or a write-once storage area to a read-only storage area. If you do not specify the Read_Only or Read_Write qualifier, the storage areas are copied with the read/write attributes that are currently in effect for the database being copied. This is a positional qualifier. 4 Read_Write Use the Read_Write qualifier to change a read-only storage area or a write-once storage area to a read/write storage area. If you do not specify the Read_Only or Read_Write qualifier, the storage areas are copied with the read/write attributes that are currently in effect for the database being copied. This is a positional qualifier. 4 Snapshots Snapshots=(Allocation=n,File=file-spec) If you specify the Allocation parameter, specifies the snapshot file allocation size in n pages for a copied area. If you specify the File parameter, specifies a new snapshot file location for the copied storage area to which it is applied. You can specify the Allocation parameter only, the File parameter only, or both parameters; however, if you specify the Snapshots qualifier, you must specify at least one parameter. The Snapshots qualifier is a positional qualifier. See the Usage Notes for information on how this qualifier interacts with the Root, File, and Directory qualifiers. 4 Spams Spams Nospams Specifies whether to enable the creation of space area management (SPAM) pages or disable the creation of SPAM pages (Nospams) for specified storage areas. This qualifier is not permitted with a storage area that has a uniform page format. When SPAM pages are disabled in a read/write storage area, the SPAM pages are initialized but they are not updated. The Spams qualifier is a positional qualifier. 4 Thresholds Thresholds=(n,n,n) Specifies new SPAM thresholds for the storage area to which it is applied (for a mixed page format storage area). The thresholds of a storage area with a uniform page format cannot be changed. See the Oracle Rdb7 Guide to Database Performance and Tuning for information on setting SPAM thresholds. The Thresholds qualifier is a positional qualifier. 3 Usage_Notes o To use the RMU Copy_Database command for a database, you must have the RMU$COPY privilege in the root file access control list (ACL) for the database to be copied or the OpenVMS SYSPRV or BYPASS privilege. o When you copy a database into a directory owned by a resource identifier, the ACE for the directory is applied to the database root file ACL first, and then the Oracle RMU ACE is added. This method is employed to prevent database users from overriding OpenVMS file security. However, this can result in a database which you consider yours, but to which you have no Oracle RMU privileges to access. See the Oracle Rdb Guide to Database Maintenance for details. o The RMU Copy_Database command provides four qualifiers, Directory, Root, File, and Snapshots, that allow you to specify the target for the copied files. The target can be just a directory, just a file name, or a directory and file name. If you use all or some of these four qualifiers, apply them as follows: - Use the Root qualifier to indicate the target for the copy of database root file. - Use local application of the File qualifier to specify the target for the copy of one or more storage areas. - Use local application of the Snapshots qualifier to specify the target for the copy of one or more snapshot files. - Use the Directory qualifier to specify a default target directory. The default target directory is the directory to which all files not qualified with the Root, File, or Snapshot qualifier are copied. It is also the default directory for files qualified with the Root, File, or Snapshot qualifier if the target for these qualifiers does not include a directory specification. Note the following when using these qualifiers: - Global application of the File qualifier when the target specification includes a file name causes Oracle RMU to copy all of the storage areas to different versions of the same file name. This creates a database that is difficult to manage. - Global application of the Snapshot qualifier when the target specification includes a file name causes Oracle RMU to copy all of the snapshot files to different versions of the same file name. This creates a database that is difficult to manage. - Specifying a file name or extension with the Directory qualifier is permitted, but causes Oracle RMU to copy all of the files (except those specified with the File or Root qualifier) to different versions of the same file name. Again, this creates a database that is difficult to manage. See Example 8. o You cannot use the RMU Copy_Database command to copy a database to a remote system or to an NFS (Network File System) mounted file system. The RMU Copy_Database command allows you to create a copy of a database on the same node as the original database. o You cannot disable extents of snapshot (.snp) files. o The file and area qualifiers for the RMU Copy_Database command are positional qualifiers, and if placed randomly, could be ignored or produce unexpected results. See the Command_ Qualifiers help entry for more information on positional qualifiers. o There are no restrictions on the use of the Nospams qualifier with mixed page format storage areas, but the use of the Nospams qualifier typically causes severe performance degradation. The Nospams qualifier is only useful where updates are rare and batched, and access is primarily by database key (dbkey). 3 Examples Example 1 The following command makes a duplicate copy of the mf_personnel database in the DISK1:[USER1] directory: $ RMU/COPY_DATABASE MF_PERSONNEL /DIRECTORY=DISK1:[USER1] Example 2 The following example shows a simple duplication of a database within a user's directory. In this instance, the duplicated database has the same content and identity as the original database. After-image journal files can be interchanged between the original and the duplicated database. Execute the RMU Dump command with the header qualifier to verify that the copied database is the same as the original database. Note that the creation date listed in the header for each database is the same. $ RMU/COPY_DATABASE MF_PERSONNEL Example 3 The following example shows a duplication of a database within a user's directory through the use of the Duplicate qualifier. In this instance, the duplicated database differs from the original database. It has the same content as the original database, but its identity is different. As a result, .aij files cannot be exchanged between the original database and the duplicate database. If you use the RMU Dump command with the header qualifier for each database, you see that the creation date for the copy and the original database is different. $ RMU/COPY_DATABASE/DUPLICATE MF_PERSONNEL Example 4 The following command copies the mf_personnel database from the DISK2:[USER2] directory to the DISK1:[USER1] directory. The Extension=Disable qualifier causes extents to be disabled for all the storage area (.rda) files in the DISK1:[USER1]mf_personnel database: $ RMU/COPY_DATABASE/EXTENSION=DISABLE/DIRECTORY=DISK1:[USER1] - _$ DISK2:[USER2]MF_PERSONNEL Example 5 The following command copies the mf_personnel database from the DISK2:[USER2] directory to the DISK2:[USER1] directory. Because the Extension=Disable qualifier is specified for only the EMPIDS_ LOW and EMPIDS_MID storage areas, extents are disabled only for those two storage area (.rda) files in the DISK2:[USER1]mf_ personnel database: $ RMU/COPY_DATABASE/DIRECTORY=DISK2:[USER1] DISK2:[USER2]MF_PERSONNEL - _$ EMPIDS_LOW/EXTENSION=DISABLE,EMPIDS_MID/EXTENSION=DISABLE Example 6 The following command uses an options file to specify that the storage area files and snapshot (.snp) files be copied to different disks. Note that storage area .snp files are located on different disks from one another and from their associated storage area (.rda) files; this is recommended for optimal performance. (This example assumes that the disks specified for each storage area file in options_file.opt are different from those where the storage area files currently reside.) $ RMU/COPY_DATABASE/OPTIONS=OPTIONS_FILE.OPT MF_PERSONNEL The options file appears as: $ TYPE OPTIONS_FILE.OPT EMPIDS_LOW /FILE=DISK1:[CORPORATE.PERSONNEL]EMPIDS_LOW.RDA - /SNAPSHOT=(FILE=DISK2:[CORPORATE.PERSONNEL]EMPIDS_LOW.SNP) EMPIDS_MID /FILE=DISK3:[CORPORATE.PERSONNEL]EMPIDS_MID.RDA - /SNAPSHOT=(FILE=DISK4:[CORPORATE.PERSONNEL]EMPIDS_MID.SNP) EMPIDS_OVER /FILE=DISK5:[CORPORATE.PERSONNEL]EMPIDS_OVER.RDA - /SNAPSHOT=(FILE=DISK6:[CORPORATE.PERSONNEL]EMPIDS_OVER.SNP) DEPARTMENTS /FILE=DISK7:[CORPORATE.PERSONNEL]DEPARTMENTS.RDA - /SNAPSHOT=(FILE=DISK8:[CORPORATE.PERSONNEL]DEPARTMENTS.SNP) SALARY_HISTORY /FILE=DISK9:[CORPORATE.PERSONNEL]SALARY_HISTORY.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]SALARY_HISTORY.SNP) JOBS /FILE=DISK7:[CORPORATE.PERSONNEL]JOBS.RDA - /SNAPSHOT=(FILE=DISK8:[CORPORATE.PERSONNEL]JOBS.SNP) EMP_INFO /FILE=DISK9:[CORPORATE.PERSONNEL]EMP_INFO.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]EMP_INFO.SNP) RESUME_LISTS /FILE=DISK11:[CORPORATE.PERSONNEL]RESUME_LISTS.RDA - /SNAPSHOT=(FILE=DISK12:[CORPORATE.PERSONNEL]RESUME_LISTS.SNP) RESUMES /FILE=DISK9:[CORPORATE.PERSONNEL]RESUMES.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]RESUMES.SNP) Example 7 The following example copies the mf_personnel database from one directory to another. In addition, by specifying the Aij_Options qualifier to add after-image journal files, it enables fixed-size journaling in the database copy and sets several of the journal options as shown in the aij_journal_options.opt file. $ RMU/COPY_DATABASE MF_PERSONNEL/DIRECTORY=DB1:[ROOT] - /AIJ_OPTIONS=AIJ_JOURNAL_OPTIONS.OPT $ TYPE AIJ_JOURNAL_OPTIONS.OPT JOURNAL IS ENABLED - RESERVE 2 - ALLOCATION IS 1024 - BACKUPS ARE MANUAL - OVERWRITE IS DISABLED - SHUTDOWN_TIMEOUT IS 120 - CACHE IS DISABLED ADD MF_PERS1 FILE DB2:[AIJONE]MF_PERS1.AIJ ADD MF_PERS2 FILE DB3:[AIJTWO]MF_PERS2.AIJ Example 8 The following example demonstrates the use of the Directory, File, and Root qualifiers. In this example: o The default directory is specified as DISK2:[DIR]. o The target directory and file name for the database root file is specified with the Root qualifier. The target directory specified with the Root qualifier overrides the default directory specified with the Directory qualifier. Thus, Oracle RMU copies the database root to DISK3:[ROOT] and names it COPYRDB.RDB. o The target directory for the EMPIDS_MID storage area is DISK4:[FILE]. Oracle RMU copies EMPIDS_MID to DISK4:[FILE]. o The target file name for the EMPIDS_LOW storage area is EMPIDS. Thus, Oracle RMU copies the EMPIDS_LOW storage area to the DISK2 default directory (specified with the Directory qualifier), and names the file EMPIDS.RDA. o The target for the EMPIDS_LOW snapshot file is DISK5:[SNAP]EMPIDS.SNP Thus, Oracle RMU copies the EMPIDS_ LOW snapshot file to DISK5:[SNAP]EMPIDS.SNP. o All the other storage area files and snapshot files in the mf_ personnel database are copied to DISK2:[DIR]; the file names for these storage areas remain unchanged. $ RMU/COPY_DATABASE DISK1:[DB]MF_PERSONNEL.RDB - _$ /DIRECTORY=DISK2:[DIR] - _$ /ROOT=DISK3:[ROOT]COPYRDB.RDB - _$ EMPIDS_MID/FILE=DISK4:[FILE], - _$ EMPIDS_LOW/FILE=EMPIDS - _$ /SNAPSHOT=(FILE=DISK5:[SNAP]EMPIDS.SNP) Example 9 The following example demonstrates how to disallow exclusive mode for a copied database. It then shows the error messages returned when a user attempts to access the copied database using the disallowed mode: $ RMU/COPY_DATABASE/TRANSACTION_MODE=NOEXCLUSIVE/DIRECTORY=[.COPY] - _$ MF_PERSONNEL.RDB %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery $ SQL SQL> ATTACH 'FILENAME mf_personnel.rdb'; SQL> SET TRANSACTION READ WRITE RESERVING EMPLOYEES FOR EXCLUSIVE WRITE; %RDB-E-BAD_TPB_CONTENT, invalid transaction parameters in the transaction parameter block (TPB) -RDMS-E-INVTRANOPT, the transaction option "EXCLUSIVE WRITE" is not allowed SQL> 2 Delete_Optimizer_Statistics Deletes records from the RDB$WORKLOAD system table. 3 Description When you enable and collect workload statistics, the system table, RDB$WORKLOAD, is created and populated. (See Collect_ Optimizer_Statistics for details.) If you are knowledgeable about the data in your database, or if workload statistics were gathered for queries that are no longer in use, you might decide that you no longer want Oracle RMU to collect statistics for particular column groups. The RMU Delete Optimizer_Statistics gives you the ability to selectively delete records for column groups in the RDB$WORKLOAD system table. When you use the RMU Delete Optimizer_Statistics command, both the optimizer statistics themselves and the reference to the column duplicity factor and the null factor are deleted from the RDB$WORKLOAD system table. If you issue an RMU Collect Optimizer_Statistics command after having issued an RMU Delete Optimizer_Statistics command, statistics for the specified column group are not collected. 3 Format (B)0RMU/Delete Optimizer_Statistics root-file-spec Command Qualifiers x Defaults x /Column_Group=(column-list) x See description /[No]Log[=file-name] x See description /Tables=(table-list) x None - Required Qualifier 3 Parameters 4 root-file-spec root-file-spec Specifies the database from which optimizer statistics are to be deleted. The default file type is .rdb. 3 Command_Qualifiers 4 Column_Group Column_Group=(column-list) Specifies a list of columns that comprise a single column group. The columns specified must be a valid column group for a table specified with the Tables=(table-list) qualifier. (Use the RMU Show Optimizer_Statistics command to display a valid column groups.) When you specify the Column_Group qualifier, the entire record in the RDB$WORKLOAD system table that holds data for the specified column group is deleted. Therefore, the next time you issue the RMU Collect Optimizer_Statistics command, statistics for the specified column-group are not collected. 4 Log Log Nolog Log=file-name Specifies whether the statistics deleted from the RDB$WORKLOAD system table are to be logged. Specify the Log qualifier to have the information displayed to SYS$OUTPUT. Specify the Log=file- spec qualifier to have the information written to a file. Specify the Nolog qualifier to prevent display of the information. If you do not specify any variation of the Log qualifier, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 4 Tables Tables=(table-list) Specifies the table or tables for which column group entries are to be deleted, as follows: o If you specify the Tables=(table-list) qualifier, but do not specify the Column_Group qualifier, then all column group entries for the listed tables are deleted from the RDB$WORKLOAD system table. o If you specify the Tables=(table-list) qualifier, and you specify the Column_Group=(column-list) qualifier, then the workload statistics entries for the specified tables that have exactly the specified column group are deleted from the RDB$WORKLOAD system table. o If you use an asterisk (*) with the Tables qualifier (Tables=*), all tables registered in the RDB$WORKLOAD table are deleted. This allows the RDB$WORKLOAD table to be purged. If you issue an RMU Collect Optimizer_Statistics command after you have deleted a workload column group from the RDB$WORKLOAD system table, those statistics are no longer collected. The Tables=(table-list) qualifier is a required qualifier; you cannot issue an RMU Delete Optimizer_Statistics command without the Tables=(table-list) qualifier. 3 Usage_Notes o To use the RMU Delete Optimizer_Statistics command for a database, you must have the RMU$ANALYZE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o Cardinality statistics are automatically maintained by Oracle Rdb. Physical storage and workload statistics are only collected when you issue an RMU Collect Optimizer_Statistics command. To get information about the usage of physical storage and workload statistics for a given query, define the RDMS$DEBUG_FLAGS logical name to be "O". For example: $ DEFINE RDMS$DEBUG_FLAGS "O" When you execute a query, if workload and physical statistics have been used in optimizing the query, you will see a line such as the following in the command output: ~O: Workload and Physical statistics used o Oracle Corporation recommends that you execute an RMU Show Optimizer_Statistics command with the Output qualifier prior to executing an RMU Delete Optimizer_Statistics command. If you accidentally delete statistics, you can replace them by issuing an RMU Insert Optimizer_Statistics command and specifying the statistical values contained in the output file. 3 Examples Example 1 The following example issues commands to do the following: 1. Display optimizer statistics for the EMPLOYEES and JOB_HISTORY tables and their indexes 2. Delete the entries for the column group (EMPLOYEE_ID, JOB_ CODE, JOB_START, JOB_END, DEPARTMENT_CODE, SUPERVISOR_ID) in JOB_HISTORY $ RMU/SHOW OPTIMIZER_STATISTICS MF_PERSONNEL.RDB - _$ /TABLES=(EMPLOYEES, JOB_HISTORY)/STATISTICS=(WORKLOAD) ----------------------------------------------------------------------- Optimizer Statistics for table : EMPLOYEES Workload Column group : EMPLOYEE_ID Duplicity factor : 1.0000000 Null factor : 0.0000000 First created time : 3-JUL-1996 10:37:36.43 Last collected time : 3-JUL-1996 10:46:10.73 Workload Column group : LAST_NAME, FIRST_NAME, MIDDLE_INITIAL, ADDRESS_DATA_1, ADDRESS_DATA_2, CITY, STATE, POSTAL_CODE, SEX, BIRTHDAY, STATUS_CODE Duplicity factor : 1.5625000 Null factor : 0.3600000 First created time : 3-JUL-1996 10:37:36.43 Last collected time : 3-JUL-1996 10:46:10.74 ---------------------------------------------------------------------- Optimizer Statistics for table : JOB_HISTORY Workload Column group : EMPLOYEE_ID Duplicity factor : 2.7400000 Null factor : 0.0000000 First created time : 3-JUL-1996 10:37:36.43 Last collected time : 3-JUL-1996 10:54:09.62 Workload Column group : EMPLOYEE_ID, JOB_CODE, JOB_START, JOB_END, DEPARTMENT_CODE, SUPERVISOR_ID Duplicity factor : 1.5930233 Null factor : 0.3649635 First created time : 3-JUL-1996 10:57:47.65 Last collected time : 3-JUL-1996 10:57:47.65 $ ! $ ! Delete one of the entries for JOB_HISTORY $ ! $ RMU/DELETE OPTIMIZER_STATISTICS MF_PERSONNEL.RDB/TABLE=(JOB_HISTORY) - _$ /COLUMN_GROUP=(EMPLOYEE_ID,JOB_CODE,JOB_START,JOB_END, - _$ DEPARTMENT_CODE,SUPERVISOR_ID)/LOG Changing RDB$SYSTEM area to READ_WRITE. Workload column group deleted for JOB_HISTORY : EMPLOYEE_ID, JOB_CODE, JOB_START, JOB_END, DEPARTMENT_CODE, SUPERVISOR_ID 2 Dump Dumps the contents of database files, including: storage area files, snapshot files, recovery-unit journal files, after-image journal files, optimized after-image journal files, and root files. You can dump database file contents to your terminal screen or to a text file. 3 Database Displays or writes to a specified output file the contents of database, storage area (.rda), and snapshot (.snp) files, including root information. NOTE The Start and End qualifiers apply only when the Areas, Lareas, Snapshots, Abms_Only or Spams_Only qualifier is specified. 4 Description Use this command to examine the contents of your database root (.rdb), storage area (.rda), and snapshot (.snp) files, to display current settings for database definition options, and to display a list of active database users. The list of database users is maintained clusterwide in a VMScluster environment. You can display the contents of all pages in any data storage area of the database or display the contents of just those pages in which rows and indexes for a specific table are stored. See the chapter that explains the internal database page format in the Oracle Rdb Guide to Database Maintenance for tutorial information. Depending on your selection of qualifiers, the RMU Dump command can list: o A formatted display of any number of pages in the storage area of the database. o A formatted display of any number of pages in a uniform logical area of the database. o A formatted display of any number of pages in the snapshot area of the database. o Header information. (This is listed by default if no qualifiers are specified.) o Current users of the database. 4 Format (B)0 RMU/Dump root-file-spec File Qualifiers x Defaults x /ABMS_Only x See description /[No]Areas [= storage-area-list] x /Noareas /End=integer x See description /[No]Header[=detail-opt, type-opts] x See description /[No]Lareas [ = logical-area-list] x /Nolareas /Option={Normal | Full | Debug} x /Option=Normal /Output = file-name x /Output=SYS$OUTPUT /Restore_Options=file-name x None /[No]Snapshots [ = storage-area-list] x /Nosnapshots /Spams_Only x See description /Start=integer x See description /State=Blocked x See description /[No]Users x /Nousers 4 Parameters 5 root-file-spec A file specification for the database root file whose root file header information, user information, storage area file pages, or snapshot area file pages you want to display. 4 Command_Qualifiers 5 ABMS_Only Specifies that the RMU/DUMP command will only dump ABM pages in uniform storage areas or in logical areas contained within uniform storage areas. The ABM pages can be dumped within a limited page range specified by the START and END qualifiers. If there are no ABM pages within the specified page range or the storage area is a mixed format area or the logical area is contained within a mixed storage area, no ABM pages will be dumped. This qualifier cannot be specified in the same Dump command as the SPAMS_Only qualifier. This qualifier cannot be specified in the same Dump command with the Snapshots qualifier. 5 Areas Areas [=storage-area-list] Noareas Specifies a display that consists of storage area pages. You can specify storage areas by name or by the area's ID number. If you specify more than one storage area, separate the storage area names or ID numbers in the storage area list with a comma, and enclose the list within parentheses. You can also specify the Areas=* qualifier to display all storage areas. If you do not specify the Areas qualifier, none of the storage areas are displayed. You can use the Start and End qualifiers to display a range of storage area pages. The Areas qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 5 End End=integer Specifies the highest-numbered area or snapshot page to include in the display. The default is the last page. If you also use the Lareas qualifier, note that the Start and End qualifiers specify a page range relative to the logical area, not a specific storage area page number. 5 Header Header Noheader Header[=(detail-opt, type-opts)] Indicates whether to include the database header in the output. Specify the Header qualifier to include all database header information in the output. Specify the Noheader qualifier to suppress the database header listing. Specify the Header=(detail- opt, type-opts) qualifier to limit the output from the header to specific items of interest. Use the detail-opt options (Brief or Detail) to limit the amount of output. Use the type-opt options to limit the output to specific types of information. RMU Dump Command Header Options summarizes the Header options and the effects of specifying each option. Table 8 RMU Dump Command Header Options Option Effect All Generates the full output of all the header information. If you specify this option and other Header options, the other options are ignored. This is the default option. Areas Output displays information about active storage areas and snapshot areas. Backup Output displays information about backup and recovery. Brief Generates a summary of the requested database root file information. Buffers Output displays information about database buffers. Corrupt_Page Output displays the Corrupt Page Table (CPT). Detail Generates a complete report of the requested database root file information. This is the default. Fast_Commit Output displays information about whether fast commit is enabled or disabled, whether commit to AIJ optimization is enabled or disabled, the AIJ checkpointing intervals, and the transaction interval. Hot_Standby Output displays information regarding hot standby databases. Locking Output displays information about database locking, such as whether or not adjustable record locking, carry-over lock optimization, and lock tree partitioning are enabled or disabled, and fanout factors. Journaling Output displays information about RUJ and AIJ journaling. Nodes Output displays names of nodes that are accessing the specified database. Parameters Output displays basic root file header information. Root_Record Output describes the Oracle Rdb specific section of the database root. This includes backup, restore, verify, and alter timestamps, as well as flags that indicate that no such operation has been performed. The bootstrap DBKEY is used to locate the RDB$DATABASE row for this database, and then the other system tables. If an alternate bootstrap DBKEY exists, then this database has been converted using RMU Convert Nocommit command. In this case, the current metadata version is displayed. Row_Caches Output displays information about row caches. Security_Audit Output displays information about security auditing. Sequence_Numbers Output displays database sequence numbers. Users Output displays information about active database users. If you specify both the Detail option and the Brief option, Detail takes precedence. If you specify the All option and other detail-opt options, the All option takes precedence. If you specify the Brief option or the Detail option only, the default for the type-opt is All. If you specify type-opts options, but do not specify a detail-opt option, the default for the detail-opt is Detail. If you specify more than one option, separate the options with commas and enclose the list within parentheses. See the Usage_Notes help entry under this command for information on understanding the derived values found in the database header. The Header=All and Header=Root_Record qualifiers output information on the use of the RMU Alter command on the specified database. For example, you see the following line in the output if you have never used the RMU Alter command on the database: Database has never been altered Do not confuse this with alterations made by SQL ALTER statements. Information about alterations made with the SQL ALTER statement is not included in the output from the RMU Dump command. If you specify the Areas, Lareas, or Snapshots qualifier, the Noheader qualifier is the default. Otherwise, Header=(All, Detail) is the default. It is invalid to specify the Header=Root_Record and the Option=Debug qualifiers in the same Oracle RMU command line. See the Oracle Rdb7 and Oracle CODASYL DBMS: Guide to Hot Standby Databases manual for information about the "Hot Standby" references in the database header. For complete information on the contents of the database header, see the Oracle Rdb Guide to Database Maintenance. 5 Lareas Lareas[=logical-area-list] Nolareas Specifies a display that consists of storage area pages allocated to a logical area or areas. In a single-file database, each table in the database is stored in its own logical area. You cannot use the Lareas qualifier with logical areas that are stored in storage areas that have a mixed page format. If you specify more than one logical area name, separate the storage area names in the logical area list with a comma, and enclose the list within parentheses. You can also specify the Lareas=* qualifier to display all logical areas that have a uniform page format. The default is the Nolareas qualifier. The Lareas qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 5 Option Option=type Specifies the type of information and level of detail the output will include. Three types of output are available: o Normal The output includes summary information. This is the default. o Full In addition to the Normal information, the output includes more detailed information. o Debug In addition to Normal and Full information, the output includes internal information about the data. In general, use the Debug option for diagnostic support purposes. 5 Output Output=file-name Specifies the name of the file where output is to be sent. The default is SYS$OUTPUT. The default output file type is .lis, if you specify a file name. 5 Restore_Options Restore_Options=file-name Generates an options file designed to be used with the Options qualifier of the RMU Restore command. The Restore_Options file is created by reading the database root file. Therefore, there is no guarantee that this options file will work with all backup files you attempt to restore with a Restore operation. For example, if areas have been added or deleted from the database since the backup file was created, there will be a mismatch between the Restore_Options file and the backup file. Similarly if the backup file was created by a backup by-area operation, the Restore_Options file may refer to areas that are not in the backup file. By default a Restore_Options file is not created. If you specify the Restore_Options qualifier and a file, but not a file extension, Oracle RMU uses an extension of .opt by default. 5 Snapshots Snapshots[=storage-area-list] Nosnapshots Specifies a display that consists of snapshot file pages. The RMU Dump command does not display snapshot pages if you omit the Snapshots qualifier or if you specify the Nosnapshots qualifier. In a single-file database, there is only one snapshot file. In a multifile database, each storage area has a corresponding snapshot file. Note that this parameter specifies the storage area name, not the snapshot file name. If you specify more than one storage area name, separate the storage area names with commas, and enclose the storage-area-list within parentheses. If you specify the Snapshots qualifier without a storage area name, information is displayed for all snapshot files. You can use the Start and End qualifiers to display a range of snapshot file pages. The default is the Nosnapshots qualifier. The Snapshots qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 5 Spams_Only Spams_Only Allows you to dump only the space area management (SPAM) pages in the selected areas and page range. A common usage for the RMU Dump command is to track down problems with storage allocation and record placement. When this qualifier is used, the SPAM pages are dumped, allowing you to locate the individual data pages that you want to examine. There is no negated form for this qualifier, and, if it is omitted, all the selected pages are dumped. The Start and End qualifiers can be used with the Spams_Only qualifier. 5 Start Start=integer Specifies the lowest-numbered area or snapshot page to include in the display. The default is the first page; that is, the Start=1 qualifier. If you also use the Lareas qualifier, note that the Start and End qualifiers specify a page range relative to the logical area, not a specific storage area page number. 5 State State=Blocked Specifies a list of all unresolved distributed transactions in the blocked database. A blocked database is a database that is not committed or rolled back and is involved in an unresolved distributed transaction. The State=Blocked qualifier displays the following information about each transaction: o Process identification (PID) o Stream identification o Monitor identification o Transaction identification o Name of the recovery journal o Transaction sequence number (TSN) o Distributed transaction identifier (TID) o Name of the node on which the failure occurred o Name of the node initiating the transaction (parent node) You can use the State=Blocked qualifier only with the Users qualifier. For information on resolving unresolved transactions with the RMU Dump command, see the Oracle Rdb7 Guide to Distributed Transactions. 5 Users Users Nousers Lists information about the current users of the database, including all users in a VMScluster environment. Oracle RMU does not consider a process that is running the Performance Monitor (with the RMU Show Statistics command or through the Windowing interface) to be a database user. The default is Nousers. 4 Usage_Notes o To use the RMU Dump command with the Areas qualifier or the Lareas qualifier or the Snapshots qualifier for a database, you must have the RMU$DUMP privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. To use the RMU Dump command with the Header qualifier for a database, you must have the RMU$DUMP, RMU$BACKUP, or RMU$OPEN privileges in the root file access control list (ACL) for the database, or the OpenVMS SYSPRV or BYPASS privilege. To use the RMU Dump command with the Users qualifier, you must have the RMU$DUMP, RMU$BACKUP, or RMU$OPEN privileges in the root file access control list (ACL) for the database or the OpenVMS WORLD privilege. o The Spams_Only qualifier conflicts with the Lareas and Snapshots qualifiers; an error is generated if you specify the Spams_Only qualifier with either of the other qualifiers. o The Header=All and Header=Buffers qualifiers provide two derived values to provide an estimated size of the global section. These appear in the dump file as: Derived Data... - Global section size With global buffers disabled is 43451 bytes With global buffers enabled is 941901 bytes The first value (With global buffers disabled) indicates the approximate size of the global section when local buffers are being used. The second value (With global buffers enabled) indicates the approximate size of the global section if you were to enable global buffers. You can use these values to determine approximately how much bigger the global section becomes if you enable global buffers. This allows you to determine, without having to take the database off line, how much larger to make the VIRTUALPAGECNT and GBLPAGES SYSGEN parameters to accommodate the larger global section. However, note that you must take the database off line if you decide to enable global buffers and you must shut down and reboot the system to change the SYSGEN parameters. It is recommended that you run AUTOGEN after you change SYSGEN parameters. Also note that these changes may require you to change the MONITOR account quotas as well to ensure the paging file quota is adequate. 4 Examples Example 1 The following example displays the header information for the mf_personnel database on the terminal screen: $ RMU/DUMP MF_PERSONNEL Example 2 The following example generates a list of unresolved transactions for the mf_personnel database: $ RMU/DUMP/USERS/STATE=BLOCKED MF_PERSONNEL Example 3 The following example shows the command you might use to view the SPAM pages associated with the area EMPIDS_LOW: $ RMU/DUMP/NOHEADER/AREAS=(EMPIDS_LOW)/SPAMS_ONLY - _$ MF_PERSONNEL/OUTPUT=DUMP.LIS Example 4 The following example demonstrates the use of the Restore_Options qualifier. The first command performs a dump operation on the mf_ personnel database and creates a Restore_Options file. The second command shows a portion of the contents of the options file. The last command demonstrates the use of the options file with the RMU Restore command. $ RMU/DUMP MF_PERSONNEL.RDB /RESTORE_OPTIONS=MF_PERS.OPT - _$ /OUTPUT=DUMP.LIS $ TYPE MF_PERS.OPT ! Options file for database USER1:[DB]MF_PERSONNEL.RDB;1 ! Created 19-JUL-1995 14:55:17.80 ! Created by DUMP command RDB$SYSTEM - /file=USER2:[STO]MF_PERS_DEFAULT.RDA;1 - /extension=ENABLED - /read_write - /spams - /snapshot=(allocation=100, - file=USER2:[SNP]MF_PERS_DEFAULT.SNP;1) DEPARTMENTS - /file=USER3:[STO]DEPARTMENTS.RDA;1 - /blocks_per_page=2 - /extension=ENABLED - /read_write - /spams - /thresholds=(70,85,95) - /snapshot=(allocation=100, - file=USER3:[SNP]DEPARTMENTS.SNP;1) . . . $ RMU/RESTORE MF_PERSONNEL.RBF/OPTIONS=MF_PERS.OPT Example 5 The following command generates a detailed display of backup, recovery, RUJ, and AIJ information for the mf_personnel database. $ RMU/DUMP/HEADER=(BACKUP,JOURNALING) MF_PERSONNEL.RDB See the Oracle Rdb Guide to Database Maintenance and the Oracle Rdb7 Guide to Distributed Transactions for more examples showing the RMU Dump command and the output. Example 6 The following example dumps all ABM pages contained in all uniform storage areas in the specified Rdb database. $ RMU/DUMP/ABMS_ONLY/OUT=DMP.OUT MF_PERSONNEL Example 7 In the following example, only the ABM pages contained in the named uniform storage area in the specified Rdb database are dumped. $ RMU/DUMP/ABMS_ONLY/AREA=RDB$SYSTEM MF_PERSONNEL Example 8 In the following example, only the ABM pages contained in the named logical area in a uniform storage area in the specified Rdb database are dumped. $ RMU/DUMP/ABMS_ONLY/LAREA=RDB$RELATIONS MF_PERSONNEL Example 9 In the following example, only the ABM pages contained within the specified page range in the named uniform storage area in the specified Rdb database are dumped. $ RMU/DUMP/ABMS_ONLY/AREA=RDB$SYSTEM/START=1/END=5 MF_PERSONNEL 3 After_journal Displays an after-image journal (.aij) file, a backed up .aij file (.aij if the backup is on disk, .aij_rbf if the .aij file was backed up to tape), or an optimized after-image journal (.oaij) file in ASCII format. Use this command to examine the contents of your .aij, .aij_rbf, or .oaij file. Whenever the term .aij file is used in this RMU Dump After_Journal command description, it refers to .oaij and .aij_rbf files, as well as .aij files. An .aij file contains header information and data blocks. Header information describes the data blocks, which contain copies of data stored in the database file. 4 Description The RMU Dump After_Journal command specifies an .aij file, not a database file, as its parameter, and is a separate command from the RMU Dump command used to display database areas and header information. The .aij file is in binary format. This command translates the binary file into an ASCII display format. The RMU Dump After_Journal command always includes the header of the .aij file in the display. You can use the Nodata qualifier to exclude data blocks from the display entirely, or you can use the Start and End qualifiers to restrict the data block display to a specific series of blocks. If you do not specify any of these qualifiers, Oracle RMU includes all data blocks. 4 Format (B)0RMU/Dump/After_Journal aij-file-name File Qualifiers x Defaults x /Active_IO=max-reads x /Active_IO=3 /Area=integer x None /[No]Data x /Data /Encrypt=({Value=|Name=}[,Algorithm=]) x See description /End=integer x See description /First=(select-list) x See description /Format={Old_File|New-Tape} x Format=Old_File /Label=(label-name-list) x See description /Larea=integer x None /Last=(select-list) x See description /Librarian[=options] x None /Line=integer x None /[No]Media_Loader x See description /Only=(select-list) x See description (B)0/Option={Statistics|Nostatistics} x Option=Statistics /Output=file-name x /Output=SYS$OUTPUT /Page=integer x None /Prompt={Automatic|Operator|Client} x See description /No]Rewind x Norewind /Start=integer x See description /State=Prepared x See description 4 Parameters 5 aij-file-name The .aij file you want to display. The default file type is .aij. For .oaij files, you must specify the file type of .oaij. 4 Command_Qualifiers 5 Active_IO Active_IO=max-reads Specifies the maximum number of read operations from a backup device that the RMU Dump After_Journal command will attempt simultaneously. This is not the maximum number of read operations in progress; that value is the product of active system I/O operations. The value of the Active_IO qualifier can range from 1 to 5. The default value is 3. Values larger than 3 can improve performance with some tape drives. 5 Area Area=integer Identifies a physical database storage area by number. Dump output is limited to the specified area. The minimum value is 1. 5 Data Data Nodata Specifies whether you want to display data blocks of the .aij file, or just the .aij file header. The Data qualifier is the default. It causes the display of the .aij file data blocks (in addition to the file header) in an ASCII display format. The Nodata qualifier limits the display to the record headers of the .aij file. 5 Encrypt Encrypt=({Value=|Name=}[,Algorithm=]) The Encrypt qualifier decrypts the file of an after-image journal backup. Specify a key value as a string or the name of a predefined key. If no algorithm name is specified the default is DESCBC. For details on the Value, Name and Algorithm parameters type HELP ENCRYPT at the OpenVMS prompt. This feature requires the OpenVMS Encrypt product to be installed and licensed on your system. This feature only works for a newer format backup file which has been created using the Format=New_Tape qualifier. You must specify the Format=New_Tape qualifier with this command if you use the Encrypt qualifier. 5 End End=integer Specifies the number of the last data block that you want to display. The default integer is the number of the last data block in the file. If you do not use the End qualifier, Oracle RMU displays the entire .aij file. 5 First First=(select-list) Allows you to specify where you want the dump output to begin. (See the Last=(select-list) qualifier for the end of the range.) If you specify more than one keyword in the select-list, separate the keywords with commas and enclose the list in parentheses. If you specify multiple items in the select list, the first occurrence is the one that will activate Oracle RMU. For example, if you specify First=(Block=100,TSN=0:52),the dump will start when either block 100 or TSN 52 is encountered. The First and Last qualifiers are optional. You can specify both, either, or neither of them. The keywords specified for the First qualifier can differ from the keywords specified for the Last qualifier. The select-list of the First qualifier consists of a list of one or more of the following keywords: o BLOCK=block-number Specifies the first block in the AIJ journal. o RECORD=record-number Specifies the first record in the AIJ journal. This is the same as the existing Start qualifier, which is still supported but obsolete. o TID=tid Specifies the first TID in the AIJ journal. o TIME=date_time Specifies the first date and time in the AIJ journal, using absolute or delta date-time format. o TSN=tsn Specifies the first TSN in the AIJ journal, using the standard [n:]m TSN format. By default, the entire .aij file is dumped. 5 Format Format=Old_File Format=New_Tape Specifies whether the backup or optimized .aij file was written in the old (disk-optimized) or the new (tape-optimized) format. If you enter the RMU Dump After_Journal command without the Format qualifier, the default is the Format=Old_Tape qualifier. You must specify the same Format qualifier as was used with the RMU Backup After_Journal command or the RMU Optimize After_ Journal command. If your .aij file resides on disk, you should use the Format=Old_File qualifier. If you specified the Format=Old_File qualifier when you optimized or backed up the .aij file to tape, you must mount the backup media by using the DCL MOUNT command before you issue the RMU Dump After_Journal command. Because the RMU Dump After_Journal command uses RMS to read the tape, the tape must be mounted as an OpenVMS volume (that is, do not specify the /FOREIGN qualifier with the MOUNT command). If you specify the Format=New_Tape qualifier, you must mount the backup media by using the DCL MOUNT /FOREIGN command before you issue the RMU Dump After_Journal command. Similarly, if you specify OpenVMS access (you do not specify the /FOREIGN qualifier on the DCL MOUNT command) although your .aij backup was created using the Format=New_Tape qualifier, you receive an RMU-F-MOUNTFOR error. The following tape qualifiers have meaning only when used in conjunction with the Format=New_Tape qualifier: Active_IO Label Rewind 5 Label Label=(label-name-list) Specifies the 1- to 6-character string with which the volumes of the backup file have been labeled. The Label qualifier is applicable only to tape volumes. You must specify one or more label names when you use the Label qualifier. You can specify a list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas and enclose the list of names within parentheses. In a normal dump after-journal operation, the Label qualifier you specify with the RMU Dump After_Journal command should be the same Label qualifier you specified with the RMU Backup After_ Journal command to back up your after-image journal file. The Label qualifier can be used with indirect file references. See Indirect-Command-Files for more information. 5 Larea Larea=integer Identifies a logical database storage area by number. Dump output is limited to the specified area. The minimum value is 0. 5 Last Last=(select-list) Allows you to specify where you want the dump output to end. (See the First=(select-list) qualifier for the beginning range.) If you specify more than one keyword in the select-list, separate the keywords with commas and enclose the list in parentheses. If you specify multiple items in the select list, the first occurrence is the one that will activate Oracle RMU. The First and Last qualifiers are optional. You can specify both, either, or neither of them. The keywords specified for the First qualifier can differ from the keywords specified for the Last qualifier. The select-list of the Last qualifier consists of a list of one or more of the following keywords: o BLOCK=block-number Specifies the last block in the AIJ journal. o RECORD=record-number Specifies the last record in the AIJ journal. This is the same as the existing End qualifier, which is still supported but obsolete. o TID=tid Specifies the last TID in the AIJ journal. o TIME=date_time Specifies the last date and time in the AIJ journal, using absolute or delta date-time format. o TSN=tsn Specifies the last TSN in the AIJ journal, using the standard [n:]m TSN format. By default, the entire .aij file is dumped. 5 Librarian Librarian=options Use the Librarian qualifier to restore files from data archiving software applications that support the Oracle Media Management interface. The file name specified on the command line identifies the stream of data to be retrieved from the Librarian utility. If you supply a device specification or a version number it will be ignored. Oracle RMU supports retrieval using the Librarian qualifier only for data that has been previously stored by Oracle RMU using the Librarian qualifer. The Librarian qualifier accepts the following options: o Trace_file=file-specification The Librarian utility writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian utility. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian utility. The pre-determined values are : - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical_name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian utility can use to specify catalogs or archives where Oracle Rdb backup files are stored, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of any Oracle RMU command that accesses the Librarian utility. The following OpenVMS logical names must be defined for use with a Librarian utility before you execute an Oracle RMU backup or restore operation. Do not use the Logical_Names option provided with the Librarian qualifier to define these logical names. o RMU$LIBRARIAN_PATH This logical name must be defined so that the shareable Librarian image can be loaded and called by Oracle RMU backup and restore operations. The translation must include the file type (for example, .exe), and must not include a version number. The shareable Librarian image must be an installed (known) image. See the Librarian utility documentation for the name and location of this image and how it should be installed. o RMU$DEBUG_SBT This logical name is not required. If it is defined, Oracle RMU will display debug tracing information messages from modules that make calls to the Librarian shareable image. You cannot use device specific qualifiers such as Rewind, Density, or Label with the Librarian qualifier because the Librarian utility handles the storage meda, not Oracle RMU. 5 Line Area=integer Identifies a database line number. Dump output is limited to the specified line. The minimum value is 0. This qualifier is intended for use during analysis or debugging. 5 Media_Loader Media_Loader Nomedia_Loader Use the Media_Loader qualifier to specify that the tape device from which the file is being read has a loader or stacker. Use the Nomedia_Loader qualifier to specify that the tape device does not have a loader or stacker. By default, if a tape device has a loader or stacker, Oracle RMU should recognize this fact. However, occasionally Oracle RMU does not recognize that a tape device has a loader or stacker. Therefore, when the first tape has been read, Oracle RMU issues a request to the operator for the next tape, instead of requesting the next tape from the loader or stacker. Similarly, sometimes Oracle RMU behaves as though a tape device has a loader or stacker when actually it does not. If you find that Oracle RMU is not recognizing that your tape device has a loader or stacker, specify the Media_Loader qualifier. If you find that Oracle RMU expects a loader or stacker when it should not, specify the Nomedia_Loader qualifier. 5 Only Only=(select-list) Allows you to specify one select list item to output. (See also the First=(select-list) and Last=(select-list) qualifiers for specifying a range.) If you specify more than one keyword in the select-list, separate the keywords with commas and enclose the list in parentheses. If you specify multiple items in the select list, the first occurrence is the one that will activate Oracle RMU. The Only qualifier is optional. The select-list of the Only qualifier consists of a list of one or more of the following keywords: o TID=tid Specifies a TID in the AIJ journal. o TSN=tsn Specifies a TSN in the AIJ journal, using the standard [n:]m TSN format. o Type=type-list Specifies the types of records to be dumped. The type-list consists of a list of one or more of the following keywords: - Ace_header Type=A records - Checkpoint Type=B records - Close Type=K records - Commit Type=C records - Data Type=D records - Group Type=G records - Information Type=N records - Open Type=O records - Optimize_information Type=I records - Prepare Type=V records - Rollback Type=R records By default, the entire .aij file is dumped. 5 Option Option=Statistics Option=Nostatistics The Option=Statistics qualifier specifies that you want Oracle RMU to include statistics on how frequently database pages are referenced by the data records in the .aij file. In addition, if the database root file is available, the output created by the Options=Statistics qualifier includes the value to specify for the Aij_Buffers qualifier of the RMU Recover command. If several .aij files will be used in your recovery operation, perform an RMU Dump After_Journal on each .aij file and add the recommended Aij_Buffer values. Use the total as the value you specify with the Aij_Buffers qualifier. See Example 2 in the Examples help entry under this command for an example using this qualifier. Note that the value recommended for the RMU Recover command's Aij_Buffers qualifier is the exact number of buffers required by the data records in the specified .aij file. If you specify fewer buffers, you may see more I/O, but you will not necessarily see performance degrade. (Performance also depends on whether asynchronous batch-writes are enabled.) Using more buffers than are recommended may result in your process doing more paging than required, and if so, performance degrades. If you specify the recommended value, note that this does not mean that no buffers are replaced during the recovery operation. The Oracle RMU buffer replacement strategy is affected by whether asynchronous prefetches and asynchronous batch-writes are enabled, and on the contents of the buffers before the recovery operation begins. If the database root file is not available, the Option=Statistics qualifier does not provide a value for the RMU Recover command's Aij_Buffers qualifier. However, it does provide the statistics on the frequency with which each page is accessed. Specify the Option=Nostatistics qualifier to suppress .aij statistics generation. The default for the RMU Dump After_Journal command is Option=Statistics. 5 Output Output=file-name Specifies the name of the file where output will be sent. The default is SYS$OUTPUT. The default file type is .lis, if you specify a file name. 5 Page Page=integer Identifies a database page number. Dump output is limited to the specified page. The minimum value is 1. This qualifier is intended for use during analysis or debugging. 5 Prompt Prompt=Automatic Prompt=Operator Prompt=Client Specifies where server prompts are to be sent. When you specify Prompt=Automatic, prompts are sent to the standard input device, and when you specify Prompt=Operator, prompts are sent to the server console. When you specify Prompt=Client, prompts are sent to the client system. 5 Rewind Rewind Norewind Specifies that the magnetic tape that contains the backup file will be rewound before processing begins. The tape is searched for the backup file starting at the beginning-of-tape (BOT). The Norewind qualifier is the default and causes a search for the backup file to be started at the current tape position. The Rewind and Norewind qualifiers are applicable only to tape devices. 5 Start Start=integer Specifies the number of the first data block that you want to display. If you do not use the Start qualifier, the display begins with the first record in the .aij file. 5 State State=Prepared Specifies a list of all records associated with unresolved transactions. For more information on listing unresolved transactions with the RMU Dump After_Journal command, see the Oracle Rdb7 Guide to Distributed Transactions. 4 Usage_Notes o The First and Last qualifiers have been added to make dumping portions of the .aij file easier. The Start and End qualifiers were intended to provide similar functionality, but are difficult to use because you seldom know, nor can you determine, the AIJ record number prior to issuing the command. o Be careful when searching for TSNs or TIDs as they are not ordered in the AIJ journal. For example, if you want to search for a specific TSN, use the Only qualifier and not the First and Last qualifiers. For example, assume the AIJ journal contains records for TSN 150, 170, and 160 (in that order). If you specify the First=TSN=160 and Last=TSN=160 qualifiers, nothing will be dumped because TSN 170 will match the Last=TSN=160 criteria. o To use the RMU Dump After_Journal command for an .aij file, you must have the RMU$DUMP privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o You receive a file access error message regarding the database's .aij file if you issue the RMU Dump After_Journal command with the active .aij file when there are active processes updating the database. To avoid the file access error message, use the RMU Close command to close the database (which stops entries to the .aij file), then issue the RMU Dump After_Journal command. o See the Oracle Rdb Guide to Database Maintenance for information on the steps Oracle RMU follows for tape label checking when you execute an RMU Dump After_Journal command using magnetic tapes. o Use of the wrong value for the Format qualifier typically results in a failure, but sometimes may produce unintelligible results. o The RMU Dump After_Journal command does not validate the file being dumped. If the file is not an .aij file or a backup of an .aij file, the RMU Dump After_Journal command produces unintelligible output. 4 Examples Example 1 The following command generates a list of records associated with unresolved transactions in the .aij file: $ RMU/DUMP/AFTER_JOURNAL/STATE=PREPARED PERSONNEL.AIJ Example 2 The following example shows the value to specify with the Aij_ Buffers qualifier along with information on how frequently each page is accessed. The output from this example shows that you should specify the Aij_Buffers=29 qualifier when you recover aij_one.aij. In addition, it shows that pages (1:623-625) were referenced 37 times which means that 8.9% of all data records in the dumped after-image journal file reference this page. $ RMU/DUMP/AFTER_JOURNAL/OPTION=STATISTICS aij_one.aij . . . Use "/AIJ_BUFFERS=29" when recovering this AIJ journal 1 recovery buffer referenced 37 times (1:623-625): 8.9% 1 recovery buffer referenced 23 times (4:23-25): 5.5% 1 recovery buffer referenced 22 times (4:5-7): 5.3% 1 recovery buffer referenced 21 times (4:44-46): 5.0% 1 recovery buffer referenced 20 times (4:50-52): 4.8% 1 recovery buffer referenced 19 times (4:41-43): 4.6% 2 recovery buffers referenced 18 times (4:38-40): 8.7% 1 recovery buffer referenced 17 times (4:17-19): 4.1% 1 recovery buffer referenced 16 times (4:29-31): 3.8% 2 recovery buffers referenced 15 times (4:35-37): 7.2% 1 recovery buffer referenced 14 times (4:2-4): 3.3% 2 recovery buffers referenced 13 times (4:11-13): 6.3% 3 recovery buffers referenced 12 times (4:8-10): 8.7% 2 recovery buffers referenced 11 times (5:2-4): 5.3% 4 recovery buffers referenced 10 times (4:14-16): 9.7% 1 recovery buffer referenced 9 times (4:47-49): 2.1% 2 recovery buffers referenced 8 times (1:617-619): 3.8% 1 recovery buffer referenced 6 times (4:20-22): 1.4% 1 recovery buffer referenced 2 times (1:503-505): 0.4% Journal effectiveness: 97.3% 175 data records 412 data modification records 423 total modification records 2 commit records 3 rollback records See the Oracle Rdb Guide to Database Maintenance and the Oracle Rdb7 Guide to Distributed Transactions for more examples of the RMU Dump After_Journal command. Example 3 The following example shows how to start a dump from Block 100 or TSN 52, whichever occurs first. $ RMU/DUMP/AFTER_JOURNAL /FIRST=(BLOCK=100,TSN=0:52) mf_personnel.aij Example 4 This example shows how to dump committed records only. $ RMU/DUMP/AFTER_JOURNAL /ONLY=(TYPE=COMMIT) mf_personnel.aij Example 5 This example shows the dump output when you specify an area, a page, and a line. RMU/DUMP/AFTER_JOURNAL/AREA=3/PAGE=560/LINE=1 mf_personnel.aij *----------------------------------------------------------------------------- * Oracle Rdb X7.1-00 3-NOV-2005 10:42:23.56 * * Dump of After Image Journal * Filename: DEVICE:[DIRECTORY]MF_PERSONNEL.AIJ;1 * *----------------------------------------------------------------------------- 2/4 TYPE=D, LENGTH=122, TAD= 3-NOV-2005 10:31:12.56, CSM=00 TID=6, TSN=0:640, AIJBL_START_FLG=01, FLUSH=00, SEQUENCE=1 MODIFY: PDBK=3:560:1, LDBID=0, PSN=0, FLAGS=00, LENGTH=84 0022 0000 line 1 (3:560:1) record type 34 00 0001 0002 Control information .... 79 bytes of static data 86726576696C6F54343631303000010D 0005 data '...00164Toliver.' 5020363431411120846E69766C410420 0015 data ' .Alvin. .A146 P' 009820876563616C50206C6C656E7261 0025 data 'arnell Place. ..' 3330484E12208B6175726F636F684307 0035 data '.Chocorua. .NH03' 20F03100630F72B31C00004D373138 0045 data '817M...³r.c.1ð ' 2/6 TYPE=D, LENGTH=224, TAD= 3-NOV-2005 10:31:12.56, CSM=00 TID=6, TSN=0:641, AIJBL_START_FLG=01, FLUSH=00, SEQUENCE=3 MODIFY: PDBK=3:560:1, LDBID=0, PSN=1, FLAGS=00, LENGTH=84 0022 0000 line 1 (3:560:1) record type 34 00 0001 0002 Control information .... 79 bytes of static data 86726576696C6F54343631303000010D 0005 data '...00164Toliver.' 5020363431411120846E69766C410420 0015 data ' .Alvin. .A146 P' 009820876563616C50206C6C656E7261 0025 data 'arnell Place. ..' 3330484E12208B6175726F636F684307 0035 data '.Chocorua. .NH03' 20F03100630F72B31C00004D373138 0045 data '817M...³r.c.1ð ' 3/9 TYPE=D, LENGTH=330, TAD= 3-NOV-2005 10:31:12.73, CSM=00 TID=6, TSN=0:642, AIJBL_START_FLG=01, FLUSH=00, SEQUENCE=5 MODIFY: PDBK=3:560:1, LDBID=0, PSN=2, FLAGS=00, LENGTH=84 0022 0000 line 1 (3:560:1) record type 34 00 0001 0002 Control information .... 79 bytes of static data 86726576696C6F54343631303000010D 0005 data '...00164Toliver.' 5020363431411120846E69766C410420 0015 data ' .Alvin. .A146 P' 009820876563616C50206C6C656E7261 0025 data 'arnell Place. ..' 3330484E12208B6175726F636F684307 0035 data '.Chocorua. .NH03' 20F03100630F72B31C00004D373138 0045 data '817M...³r.c.1ð ' Use "/AIJ_BUFFERS=3" when recovering this AIJ journal. Make sure you have enough working set and pagefile quota for the recommended number of buffers. 1 recovery buffer referenced 3 times (3:559-561): 50.0% 1 recovery buffer referenced 2 times (3:436-438): 33.3% 1 recovery buffer referenced 1 time (3:134-136): 16.6% Journal effectiveness: 54.5% 3 data records 6 data modification records 11 total modification records 3 commit records 3 Backup_File Displays or writes to a specified output file the contents of a backup file. Use this command to examine the contents of a backup (.rbf) file created by the RMU Backup command. 4 Description The RMU Dump Backup_File command reads an .rbf file and displays the contents. It uses an .rbf file, not a database file, as its parameter, and is a separate command from the RMU Dump command. The output captures unrecoverable media errors and indicates if there are unknown backup blocks on tape. This command can can be used to confirm that a backup file is formatted correctly and that the media is readable for the RMU Restore command. NOTE Successful completion of this command does not guarantee that data in a backup file is uncorrupt, nor that the backup file is complete, nor that a restore operation will succeed. Use the Root, Full, or Debug option to the Option qualifier to dump the database backup header information. The database backup header information includes the name of the backup file and the "Backup file database version". The "Backup file database version" is the version of Oracle Rdb that was executing at the time the backup file was created. The "Oracle Rdb structure level" listed in the section entitled "Database Parameters" is the currently executing version of Oracle Rdb. The backup header information is contained on the first volume of a database backup file on tape. 4 Format (B)0RMU/Dump/Backup_File backup-file-name Command Qualifiers x Defaults x /Active_IO=max-reads x /Active_IO=3 /Area=identity x None /Disk_File=[(Reader_Threads=n)] x /Disk_file=(Reader_Threads=1) /Encrypt=({Value=|Name=}[,Algorithm=]) x See description /End=integer x See description /Journal=file-name x See description /Label=(label-name-list) x See description /Librarian[=options] x None /[No]Media_Loader x See description /Options=options-list x See description /Output=file-name x /Output=SYS$OUTPUT /Process=process-list x See description /Prompt={Automatic|Operator|Client} x See description /Restore_Options=file-name x None /[No]Rewind x /Norewind /Skip=skip-list x See description /Start=integer x See description 4 Parameters 5 backup-file-spec A file specification for the backup file. The default file type is .rbf. If you use multiple tape drives, the backup-file-spec parameter must include the tape device specifications. Separate the device specifications with commas. For example: $ RMU/DUMP/BACKUP_FILE $111$MUA0:PERS_FULL.rbf,$112$MUA1: - _$ /LABEL=BACK01 When multiple volume tape files are processed, Oracle RMU dismounts and unloads all but the last volume containing the file, which is the customary practice for multiple volume tape files. See the Oracle Rdb Guide to Database Maintenance for more information on using multiple tape drives. 4 Command_Qualifiers 5 Active_IO Active_IO=max-reads Specifies the maximum number of read operations from the backup file that the RMU Dump Backup_File command will attempt simultaneously. The value of the Active_IO qualifier can range from 1 to 5. The default value is 3. Values larger than 3 might improve performance with multiple tape drives. 5 Area Area=identity Only dump the storage area identified by the specified name or ID number. The area name must be the name of a storage area in the database root file and the area ID number must be a storage area ID number in the database root file. This information is contained in the "Database Parameters:" section of the backup file which is output at the start of the dump. Snapshot areas are not contained in the backup file and cannot be specified. If this qualifier is used without the /START and /END qualifiers, all page records in the specified storage area will be output. 5 Disk_File Disk_File=[(Reader_Threads=integer)] Specifies that you want to dump a multiple disk backup file. This is a backup file that was created by the RMU Backup command with the Disk_File qualifier. The Reader_Threads keyword specifies the number of threads that Oracle RMU should use when performing a multithreaded read operation from disk files. You can specify no more than one reader thread per device specified on the command line (or in the command parameter options file). By default, one reader thread is used. This qualifier and all qualifiers that control tape operations (Label, Media_Loader, and Rewind) are mutually exclusive. 5 Encrypt Encrypt=({Value=|Name=}[,Algorithm=]) Specify a key value as a string or, the name of a predefined key. If no algorithm name is specified the default is DESCBC. For details on the Value, Name and Algorithm parameters see HELP ENCRYPT. This feature requires the OpenVMS Encrypt product to be installed and licensed on this system. 5 End End=integer Only dump pages ending with the specified page number in the specified storage area. This qualifier cannot be used unless the /AREA qualifier is also specified. If no pages are dumped, either the specified page or range of pages does not exist in the specified area in the backup file, or this qualifier has been used in the same RMU/DUMP/BACKUP command as an /OPTIONS, /SKIP or /PROCESS qualifier option that has excluded the specified page or range of pages from the dump. If this qualifier is not used with the /START qualifier, all page records in the specified storage area ending with the specified page number will be output. If both the /START and /END qualifiers are specified, the starting page number must be less than or equal to the ending page number. If the starting page number equals the ending page number only the page records for the specified page number are dumped. The block header for each block which contains at least one of the requested pages is dumped followed by the requested page records in that block. The START AREA record is dumped at the start of requested page records and the END AREA record is dumped at the end of the requested page records. By default, the database root parameters are dumped at the very start following the dump header. 5 Journal Journal=file-name Allows you improve tape performance by the dump backup file operation by specifying the journal file created by the RMU Backup command with the Journal qualifier. The RMU Backup command with the Journal qualifier creates the journal file and writes to it a description of the backup operation, including identification of the tape volumes, their contents, and the tape drive name. The RMU Dump Backup File with the Journal qualifier directs the RMU Dump Backup_File command to read the journal file and identify the tape volumes when the Label qualifier is not specified. The journal file must be the one created at the time the backup operation was performed. If the wrong journal file is supplied, an informational message is generated, and the specified journal file is not used to identify the volumes to be processed. 5 Label Label=(label-name-list) Specifies the 1- to 6-character string with which the volumes of the backup file have been labeled. The Label qualifier is applicable only to tape volumes. You must specify one or more label names when you use the Label qualifier. You can specify a list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas, and enclose the list of names within parentheses. In a normal dump backup operation, the Label qualifier you specify with the RMU Dump Backup_File command should be the same Label qualifier as you specified with the RMU Backup command that backed up your database. If no label is specified, the system will internally generate one consisting of the first six characters in the backup-file-spec parameter. See the Oracle Rdb Guide to Database Maintenance for information on tape label processing. The Label qualifier can be used with indirect file references. See Indirect-Command-Files for more information. 5 Librarian Librarian=options Use the Librarian qualifier to restore files from data archiving software applications that support the Oracle Media Management interface. The file name specified on the command line identifies the stream of data to be retrieved from the Librarian utility. If you supply a device specification or a version number it will be ignored. Oracle RMU supports retrieval using the Librarian qualifier only for data that has been previously stored by Oracle RMU using the Librarian qualifer. The Librarian qualifier accepts the following options: o Reader_Threads=n Use the Reader_Threads option to specify the number of backup data streams to read from the Librarian utility. The value of n can be from 1 to 99. The default is one reader thread. The streams are named BACKUP_FILENAME.EXT, BACKUP_FILENAME.EXT02, BACKUP_FILENAME.EXT03, up to BACKUP_FILENAME.EXT99. BACKUP_ FILENAME.EXT is the backup file name specified in the RMU Backup command. The number of reader threads specified for a database restore from the Librarian utility should be equal to or less than the number of writer threads specified for the database backup. If the number of reader threads exceeds the number of writer threads, the number of reader threads is set by Oracle RMU to be equal to the number of data streams actually stored in the Librarian utility by the backup. If the number of reader threads specified for the restore is less than the number of writer threads specified for the backup, Oracle RMU will partition the data streams among the specified reader threads so that all data streams representing the database are restored. The Volumes qualifier cannot be used with the Librarian qualifer. Oracle RMU sets the volume number to be the actual number of data streams stored in the specified Librarian utility. o Trace_file=file-specification The Librarian utility writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian utility. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian utility. The pre-determined values are : - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical_name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian utility can use to specify catalogs or archives where Oracle Rdb backup files are stored, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of any Oracle RMU command that accesses the Librarian utility. The following OpenVMS logical names must be defined for use with a Librarian utility before you execute an Oracle RMU backup or restore operation. Do not use the Logical_Names option provided with the Librarian qualifier to define these logical names. o RMU$LIBRARIAN_PATH This logical name must be defined so that the shareable Librarian image can be loaded and called by Oracle RMU backup and restore operations. The translation must include the file type (for example, .exe), and must not include a version number. The shareable Librarian image must be an installed (known) image. See the Librarian implementation documentation for the name and location of this image and how it should be installed. For a parallel RMU backup, define RMU$LIBRARIAN_ PATH as a system-wide logical name so that the multiple processes created by a parallel backup can all translate the logical. $ DEFINE /SYSTEM /EXECUTIVE_MODE - _$ RMU$LIBRARIAN_PATH librarian_shareable_image.exe o RMU$DEBUG_SBT This logical name is not required. If it is defined, Oracle RMU will display debug tracing information messages from modules that make calls to the Librarian shareable image. For a parallel RMU backup, the RMU$DEBUG_SBT logical should be defined as a system logical so that the multiple processes created by a parallel backup can all translate the logical. The following lines are from a backup plan file created by the RMU Backup/Parallel/Librarian command: Backup File = MF_PERSONNEL.RBF Style = Librarian Librarian_trace_level = # Librarian_logical_names = (- logical_name_1=equivalence_value_1, - logical_name_2=equivalence_value_2) Writer_threads = # The "Style = Librarian" entry specifies that the backup is going to a Librarian utility. The "Librarian_logical_names" entry is a list of logical names and their equivalence values. This is an optional parameter provided so that any logical names used by a particular Librarian utility can be defined as process logical names before the backup or restore operation begins. For example, some Librarian utilities provide support for logical names for specifying catalogs or debugging. You cannot use device specific qualifiers such as Rewind, Density, or Label with the Librarian qualifier because the Librarian utility handles the storage meda, not Oracle RMU. 5 Media_Loader Media_Loader Nomedia_Loader Use the Media_Loader qualifier to specify that the tape device from which the backup file is being read has a loader or stacker. Use the Nomedia_Loader qualifier to specify that the tape device does not have a loader or stacker. By default, if a tape device has a loader or stacker, Oracle RMU should recognize this fact. However, occasionally Oracle RMU does not recognize that a tape device has a loader or stacker. Therefore, when the first tape has been read, Oracle RMU issues a request to the operator for the next tape, instead of requesting the next tape from the loader or stacker. Similarly, sometimes Oracle RMU behaves as though a tape device has a loader or stacker when actually it does not. If you find that Oracle RMU is not recognizing that your tape device has a loader or stacker, specify the Media_Loader qualifier. If you find that Oracle RMU expects a loader or stacker when it should not, specify the Nomedia_Loader qualifier. 5 Options Options=options-list Specifies the type of information and level of detail the output will include. If you do not specify the Options qualifier or if you specify the Options=Normal qualifier, the backup file will be read, but dump output is not generated. This is useful for confirming that the backup file is structured correctly and the media is readable for the RMU Restore command. However, this command does not indicate if the data in a backup file is corrupted, nor does it guarantee that a restore operation will succeed. If you specify more than one option, you must separate the options with a comma, and enclose the options-list parameter within parentheses. Eight types of output are available: o Records Dumps the backup file record structure. o Blocks Dumps the backup file block structure. o Data The Data option can be used with either the Records option, the Blocks option, or both. When specified with the Records and Blocks options, the Data option dumps the contents of the backup file's records and blocks. When you do not specify the Data option, the Records and Blocks options dump the backup file's record structure and block structure only, not their contents. o Journal Dumps the contents of the journal file. Use the Journal option of the RMU Dump Backup_File command to direct Oracle RMU to dump the journal file created with the RMU Backup command with the Journal qualifier. The RMU Backup command with the Journal qualifier creates a journal file to which it writes a description of the backup operation, including identification of the tape volumes and their contents. You can use the output of the RMU Dump Backup_File with the Journal qualifier to identify the contents of each of the tapes that comprises the backup file. o Root Dumps the database root file contents as recorded in the backup file. This includes a dump of the database backup header information. o Normal The backup file will be read, but no dump output is generated. This is useful to verify the integrity of the backup file format and to detect media errors. o Full Specifying the Full option is the same as specifying the Root, Records, and Blocks options. Includes a dump of the database backup header information. The contents of the backup file's record structure and block structure are not dumped when the Full option is specified. o Debug Specifying the Debug option is the same as specifying the Root, Records, Blocks, Full, and Data options. The contents of the backup file's header, record structure, and block structure are dumped when the Debug option is specified. 5 Output Output=file-name Specifies the name of the file where output will be sent. The default is SYS$OUTPUT. The default output file type is .lis, if you specify a file name. 5 Process Process=process-list Specifies a list of keywords that determines how much of the backup file is to be dumped. If you specify more than one type of process-list option, separate the options with a comma, and enclose the process-list parameter within parentheses. You can specify the following three items in the process-list parameter: o Volumes=integer The number of volumes to dump, starting at the position specified in the Skip qualifier for volumes. This option is ignored if the backup file does not reside on tape. o Blocks=integer The number of blocks to dump, starting at the position specified in the Skip qualifier for blocks. This option is ignored if the backup file does not reside on tape. o Records=integer The number of records to dump, starting at the position specified in the Skip qualifier for records. This option is valid regardless of whether the backup file resides on tape or disk. 5 Prompt Prompt=Automatic Prompt=Operator Prompt=Client Specifies where server prompts are to be sent. When you specify Prompt=Automatic, prompts are sent to the standard input device, and when you specify Prompt=Operator, prompts are sent to the server console. When you specify Prompt=Client, prompts are sent to the client system. 5 Restore_Options Restore_Options=file-name Generates an options file designed to be used with the Options qualifier of the RMU Restore command. The Restore_Options file is created after the root information has been read from the backup file. By default, a Restore_Options file is not created. If you specify the Restore_Options qualifier and a file, but not a file extension, Oracle RMU uses an extension of .opt by default. 5 Rewind Rewind Norewind Specifies that the magnetic tape that contains the backup file will be rewound before processing begins. The Norewind qualifier is the default. The Rewind and Norewind qualifiers are applicable only to tape devices. You should use these qualifiers only when the target device is a tape device. See the Oracle Rdb Guide to Database Maintenance for information on tape label processing. 5 Skip Skip=skip-list Specifies a list of keywords that determines where the output display begins. The keywords indicate the position in the backup file from which to start the dump. If you specify more than one type of Skip position, separate the options with a comma, and enclose the skip-list parameter in parentheses. You can specify the following three items in the skip-list parameter: o Volumes=integer The number of volumes to ignore before starting. This option is ignored if the backup file does not reside on tape. o Blocks=integer The number of blocks to ignore before starting. This option is ignored if the backup file does not reside on tape. o Records=integer The number of records to ignore before starting. This option is valid regardless of whether the backup file resides on tape or disk. 5 Start Start=integer Only dump pages starting with the specified page number in the specified storage area. This qualifier cannot be used unless the /AREA qualifier is also specified. If no pages are dumped, either the specified page or range of pages does not exist in the specified area in the backup file, or this qualifier has been used in the same RMU/DUMP/BACKUP command as an /OPTIONS, /SKIP or /PROCESS qualifier option that has excluded the specified page or range of pages from the dump. If this qualifier is not used with the /END qualifier, all page records in the specified storage area starting with the specified page number will be output. If both the /START and /END qualifiers are specified, the starting page number must be less than or equal to the ending page number. If the starting page number equals the ending page number only the page records for the specified page number are dumped. The block header for each block which contains at least one of the requested pages is dumped followed by the requested page records in that block. The START AREA record is dumped at the start of requested page records and the END AREA record is dumped at the end of the requested page records. By default, the database root parameters are dumped at the very start following the dump header. 4 Usage_Notes o To use the RMU Dump Backup_File command for a database, you must have the RMU$DUMP, RMU$BACKUP, or RMU$RESTORE privileges in the root file access control list (ACL) for the database or the OpenVMS BYPASS privilege. You must also have read access to the .rbf file. o If you do not specify the Options qualifier or if you specify the Options=Normal qualifier, the backup file will be read, but dump output will not be generated. This is useful to verify the backup file integrity and to detect media errors. o See the Oracle Rdb Guide to Database Maintenance for examples that show the RMU Dump Backup_File command. 4 Examples Example 1 The following commands show the use of the Journal qualifier with the RMU Backup command and the RMU Dump After_Journal command. The first command creates a binary journal file that identifies the tapes used in the backup operation. The second command directs Oracle RMU to read the backup file (using the tapes identified in the BACKUP_JOURNAL.JNL file) to confirm that the backup file is structured correctly and the media is readable for the RMU Restore command. No dump output is generated because the Option qualifier is not specified. $ RMU/BACKUP MF_PERSONNEL.RDB - _$ $222$DUA20:[BCK]MF_PERSONNEL.RBF/LOG/JOURNAL=BACKUP_JOURNAL.JNL $ RMU/DUMP/BACKUP_FILE $222$DUA20:[BCK]MF_PERSONNEL.RBF - _$ /JOURNAL=BACKUP_JOURNAL.JNL Example 2 The following commands show the use of the Journal qualifier with the RMU Backup command and then with the RMU Dump Backup command. The first command creates a binary journal file that identifies the tapes used in the backup operation. The second command dumps the binary journal file created in the first command in ASCII format. $ RMU/BACKUP MF_PERSONNEL.RDB - _$ $222$DUA20:[BCK]MF_PERSONNEL.RBF/LOG/JOURNAL=BACKUP_JOURNAL.JNL $ RMU/DUMP/BACKUP_FILE $222$DUA20:[BCK]MF_PERSONNEL.RBF - _$ /JOURNAL=BACKUP_JOURNAL.JNL/OPTION=JOURNAL Example 3 The following example demonstrates the use of the Restore_Options qualifier. The first command performs a dump operation on the backup file of the mf_personnel database and creates a Restore_ Options file. The second command shows a portion of the contents of the options file. The last command demonstrates the use of the options file with the RMU Restore command. $ RMU/DUMP/BACKUP MFP.RBF /RESTORE_OPTIONS=MFP.OPT - _$ /OPTIONS=NORMAL/OUTPUT=DUMP.LIS $ TYPE MFP.OPT ! Options file for database DISK1:[DB]MF_PERSONNEL.RDB;1 ! Created 17-OCT-1995 13:09:57.56 ! Created by DUMP BACKUP command RDB$SYSTEM - /file=DISK2:[RDA]MF_PERS_DEFAULT.RDA;1 - /extension=ENABLED - /read_write - /spams - /snapshot=(allocation=248, - file=DISK3:[SNAP]MF_PERS_DEFAULT.SNP;1) EMPIDS_LOW - /file=DISK3:[RDA]EMPIDS_LOW.RDA;1 - /blocks_per_page=2 - /extension=ENABLED - /read_write - /spams - /thresholds=(70,85,95) - /snapshot=(allocation=10, - file=DISK4:[SNAP]EMPIDS_LOW.SNP;1) . . . $ RMU/RESTORE MFP.RBF/OPTIONS=MFP.OPT Example 4 The following example shows the dump of the page records for page 10 in storage area 4 in the MFP.RBF backup file. Since the /START and /END qualifiers both specify page 10, only the page records for that page are dumped. At the start of the dump is the dump header, followed by the database root parameters which are not shown to save space, followed by the block header, which begins with the "HEADER_SIZE" field, for the block which contains the records for page 10 in storage area 4, followed by the start area record for area 4 (REC_TYPE = 6), the data page header record (REC_TYPE = 7) for page 10, the data page data record (REC_TYPE = 8) for page 10, and ending with the end area record for area 4 (REC_TYPE = 11) which ends the dump. $ RMU/DUMP/BACKUP/AREA=4/START=10/END=10/OPTION=FULL MFP.RBF *------------------------------------------------------------------------------ * Oracle Rdb V7.2-420 11-JAN-2011 15:50:09.25 * * Dump of Database Backup Header * Backup filename: MFP.RBF * Backup file database version: 7.2 * *------------------------------------------------------------------------------ Database Parameters: . . . HEADER_SIZE = 80 OS_ID = 1024 UTILITY_ID = 722 APPLICATION_TYPE = 1 SEQUENCE_NUMBER = 22 MAJ_VER = 1 MIN_VER = 1 VOL_NUMBER = 1 BLOCK_SIZE = 32256 CRC = 0C5D3A78 NOCRC = 00 CRC_ALTERNATE = 00 BACKUP_NAME = MFP.RBF AREA_ID = 4 HIGH_PNO = 259 LOW_PNO = 1 HDR_CHECKSUM = 9B3D REC_SIZE = 2 REC_TYPE = 6 BADDATA = 00 ROOT = 00 AREA_ID = 4 LAREA_ID = 0 PNO = 0 REC_SIZE = 32 REC_TYPE = 7 BADDATA = 00 ROOT = 00 AREA_ID = 4 LAREA_ID = 0 PNO = 10 REC_SIZE = 28 REC_TYPE = 8 BADDATA = 00 ROOT = 00 AREA_ID = 4 LAREA_ID = 0 PNO = 10 REC_SIZE = 512 REC_TYPE = 11 BADDATA = 00 ROOT = 00 AREA_ID = 4 LAREA_ID = 0 PNO = 0 Example 5 The following example dumps the records for pages 10, 11 and 12 in the RDB$SYSTEM storage area in the MFP.RBF backup file. Following the block header containing the target records that starts with "HEADER_SIZE =", are the start area record for RDB$SYSTEM area 1 (REC_TYPE = 6), then the target ABM page records for pages 10, 11, and 12 (REC_TYPE = 10), and finally the end area record for area RDB$SYSTEM area 1 (REC_TYPE = 11) which ends the dump. $ RMU/DUMP/BACKUP/AREA=RDB$SYSTEM/START=10/END=12/OPTION=FULL MFP.RBF *------------------------------------------------------------------------------ * Oracle Rdb V7.2-420 14-JAN-2011 14:40:46.88 * * Dump of Database Backup Header * Backup filename: MFP.RBF * Backup file database version: 7.2 * *------------------------------------------------------------------------------ Database Parameters: . . . HEADER_SIZE = 80 OS_ID = 1024 UTILITY_ID = 722 APPLICATION_TYPE = 1 SEQUENCE_NUMBER = 1 MAJ_VER = 1 MIN_VER = 1 VOL_NUMBER = 1 BLOCK_SIZE = 32256 CRC = 8329C24B NOCRC = 00 CRC_ALTERNATE = 00 BACKUP_NAME = MFP.RBF AREA_ID = 1 HIGH_PNO = 178 LOW_PNO = 1 HDR_CHECKSUM = 40DE REC_SIZE = 2 REC_TYPE = 6 BADDATA = 00 ROOT = 00 AREA_ID = 1 LAREA_ID = 0 PNO = 0 REC_SIZE = 10 REC_TYPE = 10 BADDATA = 00 ROOT = 00 AREA_ID = 1 LAREA_ID = 3 PNO = 10 REC_SIZE = 10 REC_TYPE = 10 BADDATA = 00 ROOT = 00 AREA_ID = 1 LAREA_ID = 4 PNO = 11 REC_SIZE = 10 REC_TYPE = 10 BADDATA = 00 ROOT = 00 AREA_ID = 1 LAREA_ID = 4 PNO = 12 REC_SIZE = 512 REC_TYPE = 11 BADDATA = 00 ROOT = 00 AREA_ID = 1 LAREA_ID = 0 PNO = 0 3 Export Displays the contents of an export interchange (.rbr) file or a formatted .unl file created by the RMU Unload command. This is a useful debugging tool. 4 Format (B)0RMU/Dump/Export export_file Command Qualifiers x Defaults /[No]Data x /Data /[No]Options[=options-list] x /Nooptions /Output=file-name x /Output=SYS$OUTPUT 4 Parameters 5 export-file The .rbr file or formatted .unl file to be displayed. 4 Command_Qualifiers 5 Data Data Nodata The Data qualifier specifies that the contents of segmented strings and tables are to be displayed in hexadecimal format along with the ASCII translation. Specifying the Nodata qualifier excludes the contents of segmented strings and tables from the display and generates much less output. The default is the Data qualifier. 5 Options=option-list Options=option-list The Options qualifier allows the user to modify the output from the RMU Dump Export command. If you specify more than one option, you must separate the options with a comma and enclose the options-list parameter within parentheses. - ALLOCATION When importing databases for testing, the full allocation recorded in the interchange file is often not required. The clauses ALLOCATION and SNAPSHOT ALLOCATION are controlled by this option. The default is ALLOCATION. Use NOALLOCATION to omit these clauses from the generated SQL script. This option is ignored if NOIMPORT_DATABASE is specified or defaulted for the OPTIONS qualifier. - FILENAME_ONLY When importing databases for testing, the full file specification for the database root, storage areas and snapshot areas recorded in the interchange file is often not required. The FILENAME clauses are controlled by this option which trims the specification to only the filename portion. The default is NOFILENAME_ONLY. Use FILENAME_ONLY to truncate the file specification in the generated SQL script. This option is ignored if NOIMPORT_DATABASE is specified or defaulted for the OPTIONS qualifier. - HEADER_SECTION This option allows the database administrator to display just the header portion of the interchange file and avoid dumping the data or metadata for every row in the table. - IMPORT_DATABASE This keyword requests that the output from RMU Dump Export be formatted as a SQL IMPORT DATABASE statement. It uses the database attributes present in the interchange file formatted as SQL clauses. Of particular interest are the CREATE STORAGE AREA clauses which are required to IMPORT the source interchange (.rbr) file. The keyword HEADER_SECTION is implicitly selected when IMPORT_ DATABASE is used, limiting the I/O to the interchange file to the section containing the database attributes. The default is NOIMPORT_DATABASE. 5 Output Output=file-name Specifies the name of the file where output is sent. The default is SYS$OUTPUT. The default output file type is .lis, if you specify a file name. 4 Usage_Notes o You do not need Oracle RMU privileges to use the RMU Dump Export command. However, you must have OpenVMS read access to the .rbr or .unl file, or OpenVMS BYPASS privilege. o If the source interchange file is created by RMU Unload, then it does not contain any IMPORT DATABASE information and the generated SQL script cannot be used to create a database from such an interchange file. $ RMU/DUMP/EXPORT/OPTION=IMPORT_DATABASE EMPLOYEES.UNL/OUT=EMP.SQL $ SQL$ @EMP.SQL SQL> IMPORT DATABASE cont> from 'DISK1:[TESTING]EMPLOYEES.UNL;1' cont> -- ident ' Load/Unload utility' cont> -- backup file version 4 cont> -- database ident 'Oracle Rdb V7.2-131' cont> filename 'DB$:MF_PERSONNEL' cont> ; %SQL-F-EXTRADATA, unexpected data at the end of the .RBR file $ o The IMPORT_DATABASE option is intended to create a SQL script as an aid to the database administrator. Some editing of the generated script may be required under some circumstances. Only a subset of the database attributes are dumped by RMU for the IMPORT_DATABASE output. Continue to use the RMU Dump Export Option=NOIMPORT_DATABASE to see all attributes recorded in the interchange file. 4 Examples Example 1 The following is an example of the RMU Dump Export command using the default qualifiers: $ RMU/DUMP/EXPORT EMPLOYEES.UNL Example 2 The following is an example of how to use the HEADER_SECTION option to display just the header portion of the interchange file, and avoid dumping the data or metadata for every row in the table. $ RMU/DUMP/EXPORT/OPTION=HEADER JOBS.UNL BEGIN HEADER SECTION - (0) NONCORE_TEXT HDR_BRP_ID - (20) : Load/Unload utility CORE_NUMERIC HDR_BRPFILE_VERSION - (1) : 4 NONCORE_TEXT HDR_DBS_ID - (18) : Oracle Rdb V7.2-10 NONCORE_TEXT HDR_DB_NAME - (16) : DB$:MF_PERSONNEL NONCORE_DATE HDR_DB_LOG_BACKUP_DATE - (8) : 3-JUL-2006 16:52:32.83 CORE_NUMERIC HDR_DATA_COMPRESSION - (1) : 1 END HEADER SECTION - (0) In this example, the output describes the creator of the interchange file (RMU/UNLOAD), the version of Rdb used to create the file, the file specification of the database used, the date and time the interchange file was created, and an indication that compression was used by RMU Unload. 3 Recovery_Journal Displays a recovery-unit journal (.ruj) file in ASCII format. Use this command to examine the contents of an .ruj file. You might find .ruj files on your system following a system failure. An .ruj file contains header information and data blocks. Header information describes the data blocks, which contain copies of data modified in the database file. 4 Description The RMU Dump Recovery_Journal command specifies an .ruj file, not a database file, as its parameter, and is a separate command from the RMU Dump command used to display database areas and header information. The .ruj file is in binary format. This command translates the binary file into an ASCII display format. 4 Format (B)0 RMU/Dump/Recovery_Journal ruj-file-name Command Qualifiers x Defaults x /[No]Data x /Data /Output = file-name x /Output=SYS$OUTPUT 4 Parameters 5 ruj-file-name The .ruj file. The default file type is .ruj. 4 Command_Qualifiers 5 Data Data Nodata Specifies whether you want to display data blocks of the .ruj file or just the .ruj file header. The Data qualifier is the default. It causes the display of the .ruj file data blocks (in addition to the file header) in an ASCII display format. The Nodata qualifier limits the display to the file header of the .ruj file. 5 Output Output=file-name The name of the file where output will be sent. The default is SYS$OUTPUT. The default output file type is .lis, if you specify a file name. 4 Usage_Notes o You do not need Oracle RMU privileges to use the RMU Dump Recovery_Journal command. However, you must have OpenVMS READ access to the .ruj file or OpenVMS BYPASS privilege to use the RMU Dump Recovery_Journal command. o The RMU Dump Recovery_Journal command does not validate the file being dumped. If the file is not an .ruj file, the output from the RMU Dump Recovery_Journal command is unintelligible. o See the Oracle Rdb Guide to Database Maintenance for examples showing the RMU Dump Recovery_Journal command. 3 Row_Cache Allows you to display the in-memory contents of a row cache for an open database. 4 Description The RMU Dump Row_Cache command is intended for use as a diagnostic aid that allows you to display the in-memory contents of a row cache for an open database. Use this command to display the following information for each row in the specified cache: o GRIC - Address of the GRIC data structure for the cache slot o GRIB - Address of the GRIB data structure for the cache slot o SLOT - Slot number within the cache o NXTGRIC - Slot number of the next slot within the hash chain o LHMTE - Flag values indicating: - L - Row is latched - H - Row is marked Hot (modified since last checkpoint or sweep) - M - Row is marked Modified - T - Row is marked Too Big for (or removed from) the cache - E - End of on-disk checkpoint file; should never be seen with the RMU Dump Row_Cache command o SNAPPNO - Snapshot pointer (either snapshot page number or snapshot slot number o LEN - Length of the row in cache; 0 indicates row has been deleted o ACTLEN - Actual length of allocated space on the database page for the row o DBK - Database key for the row o REFCNT - Reference count: number of processes with this row in a cache working set o UPD_PID - Process ID of process currently updating the row in memory o RVNO - In-memory row modification count o TSN - Transaction sequence number of last transaction to modify the row 4 Format (B)0RMU/Dump/Row_Cache root-file-spec Command Qualifiers x Defaults x /Cache_Name=cachename x None /[No]Data x /Data /Output=file-name x /Output=SYS$OUTPUT 4 Parameters 5 root-file-spec Specifies the database root file for which you want to dump the row_cache contents. 4 Command_Qualifiers 5 Cache_Name Cache_Name=cachename Secifies the name of the cache you want to dump. You must specify the cache name. 5 Data Data Nodata The Data qualifier specifies that the in-memory content of a row_ cache is to be displayed in hexadecimal format along with the ASCII translation. The Data qualifier is the default. Specify the Nodata qualifier to display only header information for each cache slot. 5 Output Output=filename Specifies the name of the file where output is to be sent. The default is SYS$OUTPUT. If you specify a file name, the default output file type is .lis. 2 Extract Reads and decodes Oracle Rdb metadata and reconstructs equivalent statements in Relational Database Operator (RDO) or SQL (structured query language) code for the definition of that database. These statements can either be displayed or extracted. You can use these statements to create your database again if you no longer have the RDO or SQL code that defined your database. In addition, you can direct the RMU Extract command to produce output for the following: o An SQL or RDO IMPORT script (Items=Import) o An RMU Unload command for each table (Items=Unload) o An RMU Load command for each table (Items=Load) o An RMU Set Audit command for the database (Items=Security) o An RMU Verify command for the database (Items=Verify) 3 Description The RMU Extract command decodes information and reconstructs equivalent commands in the language you select with the Language qualifier for the definition of that database. You can extract the definitions to either a file or to SYS$OUTPUT. The RMU Extract command extracts the following character set information: o For databases: - The database default character set - The national character set o For domains: - The character set of each character data type domain - The length in characters of each character data type domain o For tables: - The character set of each character data type column - The length in characters of each character data type column The RMU Extract command may enclose object names in double quotation marks to preserve uppercase and lowercase characters, special character combinations, or reserved keywords. 3 Format (B)0RMU/Extract root-file-spec Command Qualifiers x Defaults x /Defaults[=defaults-list] x /Defaults=(quoting_rules=SQL92) /Items[=item-list] x /Items=All /Language=lang-name x /Language=SQL /[No]Log[=log-file] x /Nolog /Options=options-list x /Option=Normal /[No]Output[=out-file] x /Output=SYS$OUTPUT /Transaction_Type[= x See Description (transation_mode,options...]) x 3 Parameters 4 root-file-spec The file specification for the database root file from which you want to extract definitions. Note that you do not need to specify the file extension. If the database root file is not found, the command exits with a "file not found" error. 3 Command_Qualifiers 4 Defaults Defaults[=defaults-list] This qualifier is used to change the output of the RMU Extract command. The following defaults can be modified with the Defaults qualifier: o Allocation=integer Noallocation When you create a test database using the script generated by the RMU Extract command, the allocation from the source database may not be appropriate. You can use the Allocation keyword to specify an alternate value to be used by all storage areas, or you can use the Noallocation keyword to omit the clause from the CREATE STORAGE MAP syntax. The default behavior, when neither keyword is used, is to use the allocation recorded in the database for each storage area. See also the Snapshot_Allocation keyword. o Date_Format Nodate_Format By default, the RMU Extract process assumes that DATE types are SQL standard-compliant (that is DATE ANSI) and that the built-in function CURRENT_TIMESTAMP returns TIMESTAMP(2) values. If your environment uses DATE VMS exclusively, then you can modify the default by specifying the default DATE_ FORMAT=VMS. The legal values are described in the Oracle Rdb SQL Reference Manual in the SET DEFAULT DATE FORMAT section. The default is Date_Format=SQL92. Use Nodate_Format to omit the setting of this session attribute from the script. o Dialect Nodialect For some extracted SQL scripts the language dialect must be specified. You can use the Dialect keyword to supply a specified dialect for the script. You can find the legal values for this option in the Oracle Rdb SQL Reference Manual in the SET DIALECT section. The default is Nodialect. o Language Nolanguage The RMU Extract commmand uses the process language, that is, the translated value of SYS$LANGUAGE, or ENGLISH, for the SET LANGUAGE command. However, if the script is used on a different system then this language might not be appropriate. You can use the Language keyword to supply a specified language for the script. Legal language names are defined by the OpenVMS system logical name table; examine the logical name SYS$LANGUAGES for a current set. Use the Nolanguage keyword to omit this command from the script. o Quoting_Rules Noquoting_Rules You can use the Quoting_Rules keyword to supply a specified setting for the script. You can find the legal values for this option in the Oracle Rdb SQL Reference Manual in the SET QUOTING RULES section. The default is Quoting_Rules=SQL92. The RMU Extract command assumes that SQL keywords and names containing non-ASCII character set values are enclosed in quotation marks. o Snapshot_Allocation=integer Nosnapshot_Allocation When you create a test database from the RMU Extract output, the snapshot file allocation from the source database may not be appropriate. You can use the Snapshot_Allocation keyword to specify an alternate value to be used by all snapshot areas, or you can use the Noallocation keyword to omit the "snapshot allocation is" clause. The default behavior, when neither keyword is used, is to use the snapshot allocation stored in the database for each snapshot area. See also the Allocation keyword. 4 Items Items[=item-list] Allows you to extract and display selected definitions. Note that each of the item names can be combined to provide shorter command lines such as the following: $ RMU/EXTRACT/NOLOG/ITEMS=(ALL,NODATABASE) MF_PERSONNEL If you omit the Items qualifier from the command line or specify it without any options, the action defaults to Items=All. The following options can be specified with the Items qualifier: o All Indicates that all database items are to be extracted. This is the default and includes all items except Alter_Database, Forward_References, Import, Load, Protections, Revoke_Entry, Security, Synonyms, Unload, Verify, Volume, and Workload options. You can use either All or Noall in combination with other items to select specific output. In the following example, the Items=All option causes all the definitions except for Triggers to be extracted and displayed: $ RMU/EXTRACT/ITEMS=(ALL,NOTRIGGERS) MF_PERSONNEL The following example displays domain and table definitions. Note that the Noall option could have been omitted: $ RMU/EXTRACT/ITEMS=(NOALL, DOMAIN, TABLE) MF_PERSONNEL o Alter_Database (or Change_Database) Noalter_Database Displays the physical database after-image journal object definition. o Catalog Nocatalog Displays all contents of the catalog created for an SQL multischema database. This item is ignored if the interface is RDO. o Collating_Sequences Nocollating_Sequences Displays all the collating sequences defined for the database that you select. Note that Oracle Rdb does not save the name of the source OpenVMS National Character Set (NCS) library and the name becomes the defined logical, NCS$LIBRARY, by default. o Constraints Noconstraints By default, table and column constraints are output by the Items=Table qualifier. If you specify Item=Noconstraints, constraint information is not extracted for any table. If you specify the Language=SQL qualifier, the default is to have Item=Constraints enabled when tables are extracted. To extract all constraints as an ALTER TABLE statement, use the Item=Constraint and Option=Defer_Constraints qualifiers. To force all constraints to be defined after tables are defined, use the Item=Tables and Option=Defer_Constraints qualifiers. o Database Nodatabase Displays the database attributes and characteristics. This includes information such as the database root file name, the number of buffers, the number of users, the repository path name, and the characteristics for each storage area. If you specify RMU Extract with the Option=Nodictionary_ References qualifier, the data dictionary path name is ignored. o Domains (or Fields) Nodomains Displays the domain definitions. If the domain was originally defined using the data dictionary path name, the output definition shows this. If the Option=Nodictionary_References qualifier is specified, the data dictionary path name is ignored and the column attributes are extracted from the system tables. o Forward_References Noforward_References Queries the dependency information in the database (RDB$INTERRELATIONS) and extracts DECLARE FUNCTION and DECLARE PROCEDURE statements for only those routines that are referenced by other database objects. The default is Noforward_Reference. The Forward_References item is used in conjunction with other Item keywords, for example, /Item=(All,Forward). o Functions Nofunctions Displays external function definitions. o Import Noimport Generates an RDO or SQL IMPORT script that defines every storage area and row cache. The Language qualifier determines whether Oracle RMU generates an RDO or SQL IMPORT script (If you specify the Language=SQL or the Language=ANSI_SQL qualifier, the same SQL IMPORT script is generated.) Because the RDO interface does not accept many of the database options added to recent versions of Oracle Rdb, Oracle Corporation recommends that you specify the Language=SQL qualifier (or accept the default). The Items=Import qualifier is useful when you want to re- create a database that is the same or similar to an existing database. Editing the file generated by Oracle RMU to change allocation parameters or add storage areas and so on is easier than writing your own IMPORT script from scratch. When Oracle RMU generates the IMPORT script, it uses an interchange file name of rmuextract_rbr in the script. Therefore, you must either edit the IMPORT script generated by Oracle RMU to specify the interchange file that you want to import, or assign the logical name RMUEXTRACT_RBR to your interchange file name. (An interchange file is created by an SQL or RDO EXPORT statement.) See Example 14 in the Examples help entry under this command. o Indexes (or Indices) Noindexes Displays index definitions, including storage map information. o Load Unload Generates a DCL command procedure containing an RMU Load or RMU Unload command for each table in the database. This item must be specified explicitly, and is not included by default when you use the Items=All qualifier. Oracle RMU generates the Fields qualifier for the Load and Unload scripts when you specify the Option=Full qualifier. If you do not specify the Option=Full qualifier, the scripts are generated without the Fields qualifier. If you specify the RMU Extract command with the Item=Unload qualifier, DCL commands are added to the script to create a file with type .COLUMNS. This file defines all the unloaded columns. The file name of the .COLUMNS file is derived from the name of the extracted table. You can reference the file by using the "@" operator within the Fields qualifer for the RMU Load and RMU Unload commands. Virtual columns, AUTOMATIC or COMPUTED BY table columns, and VIEW calculated columns appear in the .COLUMNS file as comments. o Module Nomodule Displays procedure and function definitions. This item is valid only when the Language specification is SQL; it is ignored if the Language specification is RDO or ANSI_SQL. o Outlines Nooutlines Displays query outline definitions. This item is valid only when the Language specification is SQL; it is ignored if the Language specification is RDO or ANSI_SQL. o Procedures Noprocedures Extracts external procedures. o Profiles Noprofiles Displays profiles as defined by the CREATE PROFILE statement. o Protections Noprotections Displays the protection access control list (ACL) definitions. If the protections are defined using SQL ANSI semantics, they cannot be represented in RDO. In this case, the diagnostic message warns you that the protections must be extracted using the Language=SQL qualifier. If you specify Language=ANSI_SQL, a diagnostic message warns you that the ACL-style protections cannot be extracted in ANSI format. You must explicitly specify the Protections option. It is not included by default when you use the Items=All qualifier. o Revoke_Entry Norevoke_Entry Extracts a SQL or RDO script that deletes the protections from all access control lists in the database: database, table, sequences, column, module, function, and procedure. The output script contains a series of SQL REVOKE ENTRY statements (or DELETE PROTECTION statements if the language selected is RDO) that remove the access control entry for the user and all objects. o Role Norole Displays role definitions as defined by the SQL CREATE ROLE statement. In addition, any roles that have been granted are displayed as a GRANT statement. By default, roles are not extracted, nor are they included when you specify the Items=All qualifier. o Schema Noschema Displays the schema definitions for an SQL multischema database. This option is ignored if the interface is RDO. o Sequence Nosequence Displays the sequence definitions in the database that were originally defined with the SQL CREATE SEQUENCE statement. o Security Nosecurity Displays RMU Audit commands based on information in the database. This item must be specified explicitly, and is not included by default when you use the Items=All qualifier. o Storage_Maps Nostorage_Maps Displays storage map definitions, including the list (segmented string) storage map. o Synonyms Nosynonyms Generates a report of all the synonyms defined for the database. All synonyms of a database object, including synonyms of those synonyms, are grouped together. The output is ordered by creation as recorded by the RDB$CREATED column. This report is useful for viewing all synonyms or moving them to other databases. However, since synonyms refer to many different database objects, a single set of definitions is usually not adequate when defining a new database. Oracle Corporation recommends that you use the Option=Synonym qualifier in most cases. o Tables (or Relations) Notables Displays table definitions in the same order in which they were created in the database. If the table was originally defined using the data dictionary path name, that path name is used for the definition. If you specify the Option=Nodictionary_References qualifier, the data dictionary path name is ignored and the table attributes are extracted from the system tables. If Item=Noconstraints is specified, constraint information is not extracted for any table. The Items=Tables qualifier handles domains in the following ways: - The output for this item reflects the original definitions. If a column is based on a domain of a different name, the BASED ON clause is used in RDO, and the domain name is referenced by SQL. - Any columns that are based on fields in a system table are processed but generate warning messages. - Certain domains created using RDO in a relation definition cannot be extracted for RDO because it is not possible to distinguish columns defined using a shorthand method as shown in the example that follows. In this case, the column FIELD_1 becomes or is defined as a domain. DEFINE RELATION REL1. FIELD_1 DATATYPE IS TEXT SIZE 10. END. However, this type of definition in SQL causes special domains to be created with names starting with SQL$. In this case, the SQL domain is translated into the following data type: CREATE TABLE TAB1 (COLUMN_1 CHAR(10)); The output for this item also includes the table-level constraints that can be applied: PRIMARY KEY, FOREIGN KEY, NOT NULL, UNIQUE, and CHECK. In the case of the CHECK constraint, the expression might not be translated to or from RDO and SQL due to interface differences. o Triggers Notriggers Displays trigger definitions. o User Nouser Displays user definitions as defined by the SQL CREATE USER statement. In addition, if you also specify Role with the Item qualifier, then any roles that have been granted to a user are displayed as GRANT statements. By default, Users are not displayed, nor are they displayed when you specify the Items=All qualifier. o Verify Noverify Causes the generation of an optimal DCL command procedure containing multiple RMU Verify commands. Using this command procedure is equivalent to performing a full verification (RMU Verify with the All qualifier) for the database. This command procedure can be broken down further into partial command scripts to perform partial verify operations. These partial command scripts can then be submitted to different batch queues to do a full verify operation in parallel, or they can be used to spread out a full verify operation over several days by verifying a piece of the database at a time. A partitioning algorithm is a procedure to determine what portions of the database should be verified in the same command script. For example, areas with interrelations should be verified with the same partial command script. A partitioning algorithm considers the following when creating a partial command script from the equivalent RMU Verify command with the All qualifier: 1. Each storage area is assigned to a partition. 2. For each table in the database, if the table is not partitioned, the table is put in the partial command script corresponding to that storage area; otherwise, if the table is partitioned across several storage areas, the partitions corresponding to all of the storage areas are merged into one partial command script and the table is added to this partial command script. 3. For each index in the database, the process shown in step 2 is followed. 4. For an index on a table, the index and table are merged into one partial command script. The scripts of partial RMU Verify commands are written in the form of a command procedure. Each partial command script is preceded by a label of the form STREAM_n: where n is an integer greater than 1. For example, to execute the command at label STREAM_3:, invoke the command procedure by using the following syntax: $ @ STREAM_3 The resultant command procedure is set up to accept up to four parameters, P1, P2, P3, and P4, as shown in Parameters for Generated Command File. Table 9 Parameters for Generated Command File Parameter Option Description P1 Stream_n Specifies the command stream to be executed. The variable n is the "number" of the RMU Verify command stream to be executed. If omitted, all command streams are executed. P2 [No]Log Specifies whether to use the Log qualifier in the RMU Verify command line. If omitted, the DCL verify switch value is used. P3 Read_Only | Provides the RMU Verify Protected | Transaction_Type value. If omitted, Exclusive Transaction_Type = Protected is used. P4 Specifies the name of the output file for the RMU Verify Output qualifier. If omitted, Output = SYS$OUTPUT is used. o Views Noviews Displays view definitions. If the database was defined using SQL, it is possible that the view cannot be represented in RDO. In this case, the diagnostic message warns that the view definition is being ignored, and the user should use LANGUAGE=SQL to extract the view. Note the following transformations the RMU Extract command makes when it cannot precisely replicate the SQL source code: - The RMU Extract command cannot precisely replicate derived table column names or correlation names for any select expression. The RMU Extract command generates new names for correlation names (C followed by a number) and derived table column names (F followed by a number). For example, suppose you create a view, as follows: SQL> ATTACH 'FILENAME mf_personnel'; SQL> CREATE VIEW DERIVED_1 cont> (F1) AS cont> SELECT CAST(AVG(JOB_COUNT) AS INTEGER(2)) cont> FROM (SELECT EMPLOYEE_ID, COUNT (*) cont> FROM JOB_HISTORY cont> GROUP BY EMPLOYEE_ID) AS EMP_JOBS (EMPLOYEE_ID, JOB_COUNT); SQL> COMMIT; If you issue the following RMU Extract command, you receive the output shown: $ rmu/extract/item=view/option=(match:DERIVED_1%,noheader,filename_only) - mf_personnel set verify; set language ENGLISH; set default date format 'SQL92'; set quoting rules 'SQL92'; set date format DATE 001, TIME 001; attach 'filename MF_PERSONNEL'; create view DERIVED_1 (F1) as (select CAST(avg(C2.F2) AS INTEGER(2)) from (select C4.EMPLOYEE_ID, count(*) from JOB_HISTORY C4 group by C4.EMPLOYEE_ID) as C2 (F1, F2)); commit work; - The RMU Extract command cannot generate the original SQL source code for the user-supplied names of AS clauses. This is particularly apparent when the renamed select expression is referenced in an ORDER BY clause. In such a case, the RMU Extract command generates correlation names in the form RMU$EXT_n where n is a number. For example, suppose you create a view, as follows: SQL> SET QUOTING RULES 'SQL92'; SQL> CREATE DATA FILE xyz; SQL> CREATE TABLE DOCUMENT cont> (REPORT CHAR(10)); SQL> CREATE TABLE REPORTING cont> (NAME CHAR(5)); SQL> CREATE TABLE "TABLES" cont> (CODTAB CHAR(5)); SQL> CREATE VIEW VIEW_TEST cont> (CREDIT, cont> CODTAB, cont> CODMON) AS cont> SELECT cont> C1.NAME, cont> C2.CODTAB, cont> (SELECT C7.REPORT FROM DOCUMENT C7) AS COM cont> FROM REPORTING C1, "TABLES" C2 cont> ORDER BY C1.NAME ASC, C2.CODTAB ASC, COM ASC; SQL> If you issue the following RMU Extract command, you receive the output shown: $ RMU/EXTRACT/ITEM=VIEW MF_PERSONNEL.RDB . . . create view VIEW_TEST (CREDIT, CODTAB, CODMON) as select C1.NAME, C2.CODTAB, (select DOCUMENT.REPORT from DOCUMENT) AS RMU$EXT_1 from REPORTING C1, "TABLES" C2 order by C1."NAME" asc, C2.CODTAB asc, RMU$EXT_1 asc; o Volume Novolume Displays cardinality information in a PDL-formatted file for use by Oracle Expert for Rdb. This item must be specified explicitly, and is not included by default when the Items=All qualifier is used. o Workload Noworkload Generates a DCL command language script. The script is used with the RMU Insert Optimizer_Statistics command to extract the work load and statistics stored in the RDB$WORKLOAD table. The unloaded information can be applied after a new database is created using the SQL EXPORT and IMPORT statements, or it can be applied to a similar database for use by the RMU Collect Optimizer_Statistics/Statistic=Workload command. This item must be specified explicitly, and is not included by default when the Items=All qualifier is used. The default is Noworkload. You can modify the output of the Item=Workload qualifier by specifying the following keywords with the Option qualifier: o Audit_Comment Each RMU Insert Optimizer_Statistics statement is preceded by the created and altered date for the workload entry. The default is Noaudit_comment. o Filename_Only The database file specification output for the RMU Insert Optimizer_Statistics statement is abbreviated to just the filename. o Match A subset of the workload entries based on the wildcard file name is selected. 4 Language Language=lang-name Allows you to select one of the following interfaces: o SQL When you specify the Language=SQL qualifier, Oracle RMU generates the Oracle Rdb SQL dialect. The Oracle Rdb SQL dialect is a superset of SQL92 Entry level, with language elements from Intermediate and Full SQL92 levels. It also contains language elements from SQL:1999 and extensions specific to Oracle Rdb. o ANSI_SQL When you specify the Language=ANSI_SQL qualifier and specify the Option=Normal qualifier, Oracle RMU tries to generate ANSI SQL statements that conform to the ANSI X3.135-1989 SQL standard. When you specify the Language=ANSI_SQL qualifier and the Option=Full qualifier, Oracle RMU tries to generate SQL statements that conform to the current ANSI and ISO SQL database language standards. Refer to the Oracle Rdb SQL Reference Manual for more information. Regardless of the Option parameter you specify, any Oracle Rdb specific features (such as DATATRIEVE support clauses and storage maps) are omitted. o RDO When you specify the RDO language option, Oracle RMU generates RDO statements. The default is Language=SQL. The Language qualifier has no effect on the output generated by the Items=Load, Items=Unload, and Items=Verify qualifiers. This is because these qualifiers generate scripts that contain Oracle RMU commands only. 4 Log Log[=log-file] Nolog Enable or disables log output during execution of the RMU Extract command. The log includes the current version number of Oracle Rdb, and the values of the parameter and qualifiers. The default is Nolog. The default file extension is .log. If you specify Log without specifying a file name, output is sent to SYS$OUTPUT. 4 Options Options=options-list This qualifier is used to change the output of the RMU Extract command. This qualifier is not applied to output created by the Items=Unload, Items=Load, Items=Security, or the Items=Verify qualifier. The following options can be specified with the Options qualifier: o Audit_Comment Noaudit_Comment Annotates the extracted objects with the creation and last altered timestamps as well as the username of the creator. The date and time values are displayed using the current settings of SYS$LANGUAGE and LIB$DT_FORMAT. Noaudit_Comment is the default. o Cdd_Constraints Nocdd_Constraints Specifies that tables extracted by pathname include all constraints. The Option=Nocdd_Constraints qualifier is equivalent to the Option=Defer_Constraints qualifier for tables with a pathname. This option is ignored if Item=Noconstraints is specified. When you specify the Cdd_Constraints option and the Dictionary_References option, the RMU Extract command does not generate ALTER TABLE statements to add constraints, but instead assumes they will be inherited from the data dictionary. When you use the Nocdd_Constraints option and the Dictionary_ References option, the RMU Extract command generates ALTER TABLE statements to add FOREIGN KEY and CHECK constraints after all base tables have been created. o Cdd_References Nocdd_References This option is an alias for Dictionary_References. o Column_Volume Nocolumn_Volume Directs the RMU Extract command to output the table, column, and column segmented string cardinalities based on sorted indexes. Note that this qualifier must be used in combination with the Items=Volume qualifier. If the Items=Volume qualifier is omitted, cardinalities are not displayed. RMU Extract generates data of the following type: Volume for schema MF_PERSONNEL Default volatility is 5; Table WORK_STATUS all is 3; Table EMPLOYEES all is 100; Column EMPLOYEE_ID all is 100; Column LAST_NAME all is 83; . . . Table RESUMES all is 3; List RESUME Cardinality IS 3 Number of segments is 3 Average length of segments is 24; o Debug Nodebug Dumps the internal representation for SQL clauses such as AUTOMATIC AS, VALID IF, COMPUTED BY, MISSING_VALUE, DEFAULT_ VALUE, CONSTRAINTS, SQL DEFAULT, TRIGGERS, VIEWS, and STORAGE MAPS during processing. The keyword Debug cannot be specified with the keywords Normal or Full in the same Options qualifier list. o Defer_Constraints Nodefer_Constraints Forces all constraints to be defined (using an ALTER TABLE statement) after all tables have been extracted. This option is ignored if Item=Noconstraints is specified. If Option=Nodefer_Constraints is specified, all constraints are generated with the CREATE TABLE statement. If neither Option=Defer_Constraints nor Option=Nodefer_Constraints is specified, the default behavior is to generate NOT NULL, UNIQUE, and PRIMARY KEY constraints with the CREATE TABLE statement, and generate CHECK and FOREIGN KEY constraints in a subsequent ALTER TABLE statement. o Dictionary_References Nodictionary_References Directs the RMU Extract command to output definitions for domains (fields) and tables (relations) that reference data dictionary path names rather than using the information contained in the Oracle Rdb system tables. In addition to the database statements, this option also displays the data dictionary path name stored in the database. Refer to Example 8 in the Examples help entry under this command for an example of using this option. If neither the Option=Dictionary_References qualifier nor the Option=Nodictionary_References qualifier is specified, then Oracle RMU examines the RDB$RELATIONS and RDB$FIELDS system tables to determine whether or not any domains or tables refer to the data dictionary. If references are made to the data dictionary, then the Option=Dictionary_References qualifier is the default. Otherwise, it is assumed that the data dictionary is not used, and the default is the Option=Nodictionary_ References qualifier. The Nodictionary_References keyword causes all references to the data dictionary to be omitted from the output. This is desirable if the database definition is to be used on a system without the data dictionary or in a testing environment. If the Items=Database and Option=Nodictionary_References qualifiers are selected, the data dictionary path name stored in the system table is ignored. For SQL, the NO PATHNAME clause is generated, and for RDO, the clause DICTIONARY IS NOT USED is generated. If the Items qualifier specifies Domain or Table, and the Option qualifier specifies Nodictionary_References, the output definition includes all attributes stored in the system tables. o Disable_Objects Nodisable_Objects Requests that all disabled objects be written to the RMU Extract output file as disabled (see the description for the Omit_Disabled option). Disable_Objects is the default. The Nodisable_Objects option displays the objects but omits the disabling syntax. o Domains Nodomains The Nodomains option is used to eliminate the domain name from within metadata objects. The domain name is replaced by the underlying data type. This option is designed for use with tools that do not recognize this SQL:1999 SQL language feature. Effect on /Language=SQL output: The default is Option=Domains. A SQL script generated when Option=Nodomains was specified does not include the domain name in the CREATE TABLE column definition, CREATE FUNCTION or CREATE PROCEDURE parameter definitions, or any value expression which uses the CAST function to convert an expression to a domain data type (such as the CREATE VIEW and CREATE TRIGGER statements). The output generated by the RMU Extract command for functions and procedures in the CREATE MODULE statement is not affected by the Option=Nodomains option because it is based on the original source SQL for the routine body which is not edited by the RMU Extract command. Effect on /Language=ANSI_SQL output: The default is Option=Nodomains when the Option=Normal qualifier is specified or is the default. The RMU Extract command does not generate a list of domain definitions even if the Items=Domains or Items=All qualifier is used. If you want the generated script to include a list of domain definitions, use the Options=Domains qualifier: $RMU/EXTRACT/LANGUAGE=ANSI_SQL/OPTION=DOMAINS databasename Use the Option=Full qualifier to have the use of domains included in the syntax generated for SQL:1999. o Filename_Only Nofilename_Only Causes all file specifications extracted from the database to be truncated to only the file name. The use of this qualifier allows for easier relocation of the new database when you execute the created procedure. o Full Nofull Specifies that if metadata that cannot be translated from the language that defined the database to the equivalent construct in the language specified with the Language qualifier (for example, DEFAULT for SQL and the language selected was RDO) then the metadata is displayed in comments, or Oracle RMU attempts to create a translation that most closely approximates the original construct. Nofull is identical to the Normal option. o Group_Table Nogroup_Table Specifies that indexes and storage maps are to be extracted and grouped by table. The table is extracted first, than any PLACEMENT VIA index, then any storage map, and finally all other indexes. When the Group_Table qualifier is combined with the Option=Match qualifier, you can select a table and its related storage map and indexes. The default behavior is Nogroup_Table, which means that items are extracted and grouped by type. o Header Noheader Specifies that the script header and section headers are included in the extract. This is the default. Because the header has an included date, specifying Noheader to suppress the header may allow easier comparison with other database extractions when you use the OpenVMS DIFFERENCES command. o Limit_Volume=nn Nolimit_Volume Specifies the maximum amount of data to be scanned for segmented fields. The RMU Extract command stops scanning when the limit nn is reached. The number of segments and average length of segments are calculated from the data that was scanned. Limit_Volume=1000 is the default. Nolimit_Volume specifies that a full scan for segmented strings should be done. o Match:match-string The Match option allows selection of wildcard object names from the database. The match string can contain the standard SQL wildcard characters: the percent sign (%) to match any number of characters, and the underscore (_) to match a single character. In addition, the backslash (\) can be used to prefix these wildcards to prevent them from being used in matching. If you are matching a literal backslash, use the backslash twice, as shown in the following example: Option=Match:"A1\\A2%" The match string defaults to the percent sign (%) so that all objects are selected. To select those objects that start with JOB, use the qualifier Option=Match:"JOB%". From the mf_personnel database, this command displays the definitions for the domains JOB_CODE_DOM and JOB_TITLE_DOM, the tables JOBS and JOB_HISTORY, the index JOB_HISTORY_HASH, and the storage maps JOBS_MAP and JOB_HISTORY_MAP. The match string can be quoted as shown if the string contains spaces or other punctuation characters used by DCL or other command language interfaces. Most object names are space filled; therefore, follow the match string with the percent sign (%) to match all trailing spaces. The Match option can be used in conjunction with the Item qualifier to extract specific tables, indexes, and so on, based on their name and type. If Group_Table is specified, the match name is assumed to match a table name; all indexes for that table will be extracted when the Items=Index qualifier is specified. o Multischema Nomultischema Displays the SQL multischema names of database objects. It is ignored by the Relational Database Operator (RDO). The Nomultischema option displays only the SQL single-schema names of database objects. o Normal Nonormal Includes only the specific source language code used to define the database. This is the default. In addition, this option propagates RDO VALID IF clauses as column CHECK constraints with the attribute NOT DEFERRABLE when the Language specification is SQL or ANSI_SQL. When an RDO VALID IF clause is converted, Oracle RMU generates error messages similar to the following in your log file: %RMU-W-UNSVALIDIF, VALID IF clause not supported in SQL - ignored for DEGREE. %RMU-I-COLVALIDIF, changed VALID IF clause on domain DEGREE to column check constraint for DEGREES.DEGREE The first message is a warning that the VALID IF clause could not be added to the domain definition because the VALID IF clause is not supported by SQL. The second message is an informational message that tells you the VALID IF clause was changed to a column check constraint. o Omit_Disabled Noomit_Disabled Causes all disabled objects to be omitted from the output of the RMU Extract command. This includes indexes that have MAINTENANCE IS DISABLED, USERS with ACCOUNT LOCK, and disabled triggers and constraints. The Noomit_Disabled option causes all disabled objects to be included in the output from the RMU Extract command. Noomit_ Disabled is the default. o Order_By_Name Noorder_By_Name Order_by_Name displays the storage area, cache, and journal names for the items Database, Alter_Database (also known as Change_Database), and Import in alphabetic order by the ASCII collating sequence. Noorder_By_Name displays the storage area, cache, and journal names for the items Database, Alter_Database, and Import in approximate definition order. The default ordering is approximate because a DROP STORAGE AREA, DROP CACHE, or DROP JOURNAL statement frees a slot that can be reused, thus changing the order. Noorder_By_Name is the default. You can use the logical name RDMS$BIND_SORT_WORKFILES to allocate work files, if needed. NOTE If the identifier character set for the database is not MCS or ASCII, then this option is ignored. Characters from other character sets do not sort appropriately under the ASCII collating sequence. o Synonyms Nosynonyms Causes the synonyms to be extracted immediately after the referenced object, as shown in the following excerpt from an output file created using the Item=Table qualifier: create table HISTORICAL_JOB_INFORMATION ( EMPLOYEE_ID INTEGER, USER_ID CHAR (15), JOB_TITLE TITLE, START_DATE DATE, CURRENT_SALARY MONEY_IN_DOLLARS default NULL); create synonym JOBHIST for table HISTORICAL_JOB_INFORMATION; Because synonyms can be referenced from almost any database object, if you keep the definitions close to the target object you can eliminate occurrences of undefined symbols during script execution. The default is Option=Synonyms. Use the Option=Nosynonyms qualifier to disable the display of CREATE SYNONYM statements. The synonyms referenced in database objects such as module, procedure, trigger, and table definitions are still extracted. o Volume_Scan Novolume_scan Directs the RMU Extract command to perform queries to calculate the cardinality of each table, if both the Items=Volume and Options=Volume_Scan qualifiers are specified. The default is Options=Novolume_Scan, in which case the approximate cardinalities are read from the RDB$RELATIONS system table. The Options=Volume_Scan option is ignored if the Items=Volume qualifier is not selected. o Width=n Specifies the width of the output files. You can select values from 60 to 512 characters. The default of 80 characters is appropriate for most applications. 4 Output Output=[out-file] Nooutput Names the file to which the RMU Extract command writes the data definition language (DDL) statements. The file extension defaults to .rdo, if you specify the Language=RDO qualifier; .sql, if you specify either the Language=SQL or the Language=ANSI_SQL qualifier. If you specify the Volume option only, the output file type defaults to .pdl. If you specify Load, Security, Verify, or Unload only, the output file type defaults to .com. The default is SYS$OUTPUT. If you disable the output by using the Nooutput qualifier, command scripts are not written to an output file. The Log output can be used to determine which features used by the database cannot be converted to SQL. Using Qualifiers to Determine Output Selection shows the effects of the various combinations of the Language and Options qualifiers. Table 10 Using Qualifiers to Determine Output Selection Language Option Effect on Output RDO Normal Generates RDO syntax. Full Generates RDO syntax. Dictionary_ Outputs path name references to the References repository. Nodictionary_ Converts path name references to References the repository to RDO syntax. Multischema Ignored by RDO. SQL Normal Generates SQL syntax. Full Tries to convert RDO specific features to SQL (for example, the VALID IF clause). Dictionary_ Outputs path name references to the References data dictionary. Nodictionary_ Converts path name references to References the data dictionary to SQL syntax. Multischema Selects SQL multischema naming of objects. ANSI_ Normal Generates ANSI/ISO syntax. SQL Full Generates ANSI/ISO SQL92 syntax supported by SQL. Dictionary_ Ignored for ANSI_SQL. References Nodictionary_ Converts path name references to References the data dictionary to SQL syntax. This is the default for ANSI_SQL. Multischema Selects SQL multischema naming of objects. Any Audit_Comment Adds a comment before each definition. Debug Annotates output where possible. Domains Replaces domain names for CAST expression, column and parameter definitions, and returns clauses with SQL data type. Filename_Only Truncates all file specifications extracted from the database to only the file name. Volume_Scan Forces a true count of Tables. Only valid for Items=Volume. 4 Transaction_Type Transaction_Type[=(transaction_mode,options,...)] Allows you to specify the transaction mode, isolation level, and wait behavior for transactions. Use one of the following keywords to control the transaction mode: o Automatic When Transaction_Type=Automatic is specified, the transaction type depends on the current database settings for snapshots (enabled, deferred, or disabled), transaction modes available to the process, and the standby status of the database. Automatic mode is the default. o Read_Only Starts a READ ONLY transaction. o Write Starts a READ WRITE transaction. Use one of the following options with the keyword Isolation_ Level=[level] to specify the transaction isolation level: o Read_Committed o Repeatable_Read o Serializable. Serializable is the default setting. Refer to the SET TRANSACTION statement in the Oracle Rdb SQL Reference Manual for a complete description of the transaction isolation levels. Specify the wait setting by using one of the following keywords: o Wait Waits indefinitely for a locked resource to become available. Wait is the default behavior. o Wait=n The value you supply for n is the transaction lock timeout interval. When you supply this value, Oracle Rdb waits n seconds before aborting the wait and the RMU Extract session. Specifying a wait timeout interval of zero is equivalent to specifying Nowait. o Nowait Will not wait for a locked resource to become available. 3 Usage_Notes o To use the RMU Extract command for a database, you must have the RMU$UNLOAD privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o For tutorial information on using output from the RMU Extract command to load or unload a database, refer to the Oracle Rdb Guide to Database Design and Definition. o Included in the output from the RMU Extract command is the SQL SET DEFAULT DATE FORMAT statement. This SQL statement determines whether columns with the DATE data type or CURRENT_ TIMESTAMP built-in function are interpreted as OpenVMS or SQL92 format. The RMU Extract command always sets the default to SQL92. The SQL92 format DATE and CURRENT_TIMESTAMP contain only the YEAR to DAY fields. The OpenVMS format DATE and CURRENT_TIMESTAMP contain YEAR to SECOND fields. If your database was defined with OpenVMS format DATE and CURRENT_TIMESTAMP, the default SQL SET DEFAULT DATE FORMAT 'SQL92' in the RMU Extract output causes errors to be returned when you attempt to execute that output. For example, when you define a trigger: SQL> CREATE TRIGGER SALARY_HISTORY_CASCADE_UPDATE cont> AFTER UPDATE OF JOB_CODE ON JOB_HISTORY cont> (UPDATE SALARY_HISTORY SH cont> SET SALARY_START = CURRENT_TIMESTAMP cont> WHERE (SH.EMPLOYEE_ID = JOB_HISTORY.EMPLOYEE_ID) cont> ) for each row; %SQL-F-UNSDATASS, Unsupported date/time assignment from to SALARY_START You can avoid these errors by editing the output from the RMU Extract command. Replace the SET DEFAULT DATE FORMAT 'SQL92' statement with SET DEFAULT DATE FORMAT 'VMS'. If the problem occurs in trigger definitions, you can use the CAST function instead. Specify CAST(CURRENT_TIMESTAMP AS DATE VMS) with each trigger definition that references CURRENT_TIMESTAMP. (You cannot use the CAST function within the DEFAULT clause of an SQL CREATE statement). o The following list contains a description of what the RMU Extract command generates when it encounters certain RDO statements: - RDO and the data dictionary have the concept of validation clauses at the domain level. The ANSI/ISO SQL92 standard allows CHECK constraints defined on domains. While the actions of the ANSI/ISO CHECK constraint do differ from VALID IF in some respects, the RMU Extract command extracts the VALID IF clauses as domain CHECK constraints if you specify the Language=SQL and Option=Full qualifiers. - RDO multiline descriptions Because the RDO interface removes blank lines in multiline descriptions, the description saved in the metadata is not identical to that entered by the database definition. The RMU Extract command therefore cannot completely reconstruct the original description. - Some RDO trigger definitions RDO trigger definitions that contain a trigger action within a join of two or more tables generates invalid SQL syntax. For example, the following RDO trigger definition includes a join with an embedded ERASE statement. When the RMU Extract command encounters this statement, Oracle RMU generates the invalid SQL trigger definition shown. DEFINE TRIGGER EXAMPLE AFTER ERASE FOR C1 IN EMPLOYEES EXECUTE FOR C2 IN JOB_HISTORY CROSS C3 IN EMPLOYEES WITH (((C2.EMPLOYEE_ID = C3.EMPLOYEE_ID) AND (C2.JOB_END MISSING)) AND (C3.EMPLOYEE_ID = C2.EMPLOYEE_ID)) ERASE C2 END_FOR FOR EACH RECORD. CREATE TRIGGER EXAMPLE AFTER DELETE ON EMPLOYEES (DELETE FROM JOB_HISTORY C2, EMPLOYEES C3 WHERE (((C2.EMPLOYEE_ID = C3.EMPLOYEE_ID) AND (C2.JOB_END IS NULL)) AND (C3.EMPLOYEE_ID = C2.EMPLOYEE_ID)) ) FOR EACH ROW; Note that in Oracle Rdb Version 4.1 and higher, including a trigger action within a join of two or more tables is invalid RDO syntax. For more information on this RDO restriction, see the ERASE and MODIFY entries in RDO HELP. o Oracle CDD/Repository Version 5.3 and higher support table and column constraint definition and maintenance through CDO. The RMU Extract command, by default, assumes all constraint maintenance is with SQL and so follows each CREATE TABLE with an ALTER TABLE FROM pathname to add the constraints. However, this is no longer necessary if you are using the later versions of Oracle CDD/Repository. To disable the output of the SQL ALTER TABLE statements which add constraints use the Option=Cdd_Constraint qualifier. o If the Transaction_Type qualifier is omitted from the RMU Extract command line, a READ ONLY transaction is started against the database. This behavior is provided for backward compatibility with prior Oracle Rdb releases. If the Transaction_Type qualifier is specified without a transaction mode, the default value Automatic is used. o If the database has snapshots disabled and the Transaction_ Type qualifier was omitted, the transaction is restarted as READ WRITE ISOLATION LEVEL READ COMMITTED to reduce the number of rows locked by operations performed with the Option=Volume_ Scan qualifier enabled. o When Transaction_Type=Write is specified, the RMU Extract process does not attempt to write to the database tables. o In previous versions, Oracle Rdb used derived column names based on position, for example, F1, F2. With release 7.0.6.4 and later, Oracle Rdb promotes the column names from the base table into the derived column name list. The result is a more readable representation of the view or trigger definition. In the following example the column name EMPLOYEE_ID is propagated through the derived table. In previous releases this would be named using a generic label F1. create view SAMPLE_V (EMPLOYEE_ID, COUNTS) as select C1.EMPLOYEE_ID, C1.F2 from (select C2.EMPLOYEE_ID, (select count(*) from SALARY_HISTORY C3 where (C3.EMPLOYEE_ID = C2.EMPLOYEE_ID)) from JOB_HISTORY C2) as C1 ( EMPLOYEE_ID, F2 ) order by C1.F2 asc; o The following list shows the equivalent SQL expressions matched by the RMU Extract process: - NULLIF (a, b) is eqivalent to CASE WHEN a = b THEN NULL ELSE a END - NVL (a, ..., b) or COALESCE (a, ..., b) is equivalent to CASE WHEN a IS NOT NULL THEN a ... ELSE b END - The simple CASE expression CASE a WHEN b THEN v1 WHEN NULL THEN v2 ... ELSE v3 END is equivalent to CASE WHEN a = b THEN v1 WHEN a IS NULL THEN v2 ... ELSE v3 END The RMU Extract procedure tries to decode the internal representation to as compact a SQL expression as possible. o The RMU Extract procedure decodes case expressions into ABS (Absolute) functions: ABS(a) is equivalent to: CASE WHEN a < 0 THEN -a ELSE a END In addition, similar forms of CASE expression are also converted to ABS: CASE WHEN a <= 0 THEN -a ELSE a END CASE WHEN a > 0 THEN a ELSE -a END CASE WHEN a >= 0 THEN a ELSE -a END It is possible that the RMU Extract process will change existing CASE expressions into this more compact syntax, even if they were not originally coded as an ABS function call. o If the Group_Table option is used and the Item qualifier omits one or more of the Table, Index, or Storage_Map keywords, only the included items are displayed. For example, to extract just the indexes for the EMPLOYEES table: $ RMU/EXTRACT/ITEM=INDEX/OPTION=(GROUP_TABLE,MATCH=EMPLOYEES%) To extract only the storage map and indexes for a table, use the following command: $ RMU/EXTRACT/ITEM=(STORAGE_MAP,INDEX)/OPTION=(GROUP_TABLE, - _$ MATCH=EMPLOYEES%) o If the name of the LIST storage map is not known, it can be extracted using the following generic command: $ RMU/EXTRACT/ITEM=STORAGE_MAP/OPTION=(GROUP_TABLE, - _$ MATCH=RDB$SEGMENTED_STRING%) 3 Examples Example 1 The following command extracts these database items: COLLATING_SEQUENCES, DOMAINS, TABLES, INDEXES, STORAGE_MAPS, VIEWS, SEQUENCES, and TRIGGERS. The All option is the default. The All or Noall option can be used in conjunction with other items to select specific output. For example, the Items=(All,Nodatabase) qualifier selects all metadata items except the physical database characteristics. $ RMU/EXTRACT/ITEM=(ALL, NODATABASE) MF_PERSONNEL Example 2 The following command generates a DCL command procedure containing an RMU Load command for each table in the database: $ RMU/EXTRACT/ITEMS=LOAD MF_PERSONNEL Example 3 The following command displays the protection access control list (ACL) definitions in the mf_personnel.rdb database: $ RMU/EXTRACT/ITEMS=PROTECTIONS MF_PERSONNEL.RDB Example 4 The following command generates a DCL command procedure containing an RMU Unload command for each table in the database: $ RMU/EXTRACT/ITEMS=UNLOAD MF_PERSONNEL.RDB Example 5 The following example displays index definitions: $ RMU/EXTRACT/ITEMS=INDEXES MF_PERSONNEL Example 6 The following example displays domain and table definitions. Note that the Noall option could have been omitted. $ RMU/EXTRACT/ITEMS=(NOALL,DOMAINS,TABLES) MF_PERSONNEL Example 7 The following example displays definitions for domains (fields) and tables (relations) that reference data dictionary path names rather than using the information contained in the Oracle Rdb system tables. In addition to the database statements, it also references the data dictionary path name stored in the database, as shown in the following example: $ RMU/EXTRACT/LANG=SQL/ITEM=ALL/OPTION=DIC/OUTPUT=CDD_MODEL.LOG/LOG= - _$ CDD_EXTRACT.LOG CDD_SQL_DB Example 8 The following example creates a command procedure containing a script of partial RMU Verify commands or verify command partitions for the mf_personnel database. This command procedure was created with the following RMU Extract command: $ RMU/EXTRACT/ITEM=VERIFY MF_PERSONNEL Example 9 The following command displays a query outline definition that was previously added to the mf_personnel database: $ RMU/EXTRACT/ITEMS=(OUTLINES) MF_PERSONNEL Example 10 The following command displays the after-image journal (.aij) file configuration for mf_personnel: $ RMU/EXTRACT/ITEMS=(ALTER_DATABASE) MF_PERSONNEL Example 11 The following command displays the function definitions in mf_ personnel for functions previously created using SQL: $ RMU/EXTRACT/ITEM=FUNCTION MF_PERSONNEL Example 12 The following command displays the table and column cardinalities based on sorted indexes: $ RMU/EXTRACT/OPTION=COLUMN_VOLUME/ITEM=VOLUME MF_PERSONNEL Example 13 The following example: o Executes an SQL EXPORT statement to create an interchange file. o Executes an RMU Extract command with the Item=Import qualifier to generate an Import script. In addition, the Option=Filename_Only qualifier is specified to prevent full file specifications from appearing in the SQL IMPORT script. (If full file specifications are used, you cannot test the script without replacing the database that was exported.) o Defines a logical to define the interchange file name used in the Import script file. o Executes the Import script file. SQL> -- Create interchange file, SAVED_PERS.RBR. SQL> -- SQL> EXPORT DATABASE FILENAME MF_PERSONNEL.RDB INTO SAVED_PERS.RBR; SQL> EXIT; $ ! $ RMU/EXTRACT/ITEM=IMPORT/OPTION=FILENAME_ONLY/OUTPUT=IMPORT_PERS.SQL - _$ MF_PERSONNEL $ DEFINE/USER RMUEXTRACT_RBR SAVED_PERS.RBR $ ! $ SQL$ SQL> @IMPORT_PERS.SQL SQL> set language ENGLISH; SQL> set default date format 'SQL92'; SQL> set quoting rules 'SQL92'; SQL> set date format DATE 001, TIME 001; SQL> SQL> -- RMU/EXTRACT for Oracle Rdb V7.2-00 2-JAN-2006 15:34:38.63 SQL> -- SQL> -- Physical Database Definition SQL> -- SQL> ----------------------------------------------------------------- SQL> import database from rmuextract_rbr cont> filename 'MF_PERSONNEL' . . . Example 14 The following example shows an extract from the generated script when the SYS$LANGUAGE and LIB$DT_FORMAT symbols are defined. The language and format will default to ENGLISH and the standard OpenVMS format if these logical names are not defined. $ DEFINE LIB$DT_FORMAT LIB$DATE_FORMAT_002,LIB$TIME_FORMAT_001 $ DEFINE SYS$LANGUAGE french $ RMU/EXTRACT/OUT=SYS$OUTPUT/ITEM=DOMAIN MF_PERSONNEL/OPT=AUDIT_COMMENT . . . -- Created on 8 janvier 2006 13:01:31.20 -- Never altered -- Created by RDB_EXECUTE -- SQL> CREATE DOMAIN ADDRESS_DATA_1 cont> CHAR (25) cont> comment on domain ADDRESS_DATA_1 is cont> ' Street name'; . . . Example 15 If a database has snapshots set to ENABLED DEFERRED, it may be preferable to start a read/write transaction. In this environment, using the Transaction_type=(Read_only) qualifier causes a switch to a temporary snapshots ENABLED IMMEDIATE state. This transition forces the READ ONLY transaction to wait while all READ WRITE transactions complete, and then all new READ WRITE transactions performing updates will start writing rows to the snapshot files for use by possible read only transactions. To avoid this problem use an RMU Extract command specifying a READ WRITE ISOLATION LEVEL READ COMMITTED transaction. $ RMU/EXTRACT/ITEM=TABLE/OUT=TABLES.SQL- /TRANSACTION_TYPE=(WRITE,ISOLATION=READ)- SAMPLE.RDB Example 16 This example specifies the options which were the default transaction style in prior releases. $ RMU/EXTRACT/ITEM=TABLE/OUT=TABLES.SQL- /TRANSACTION_TYPE=(READ_ONLY)- SAMPLE.RDB Example 17 If the database currently has snapshots deferred, it may be more efficient to start a read-write transaction with isolation level read committed. This allows the transaction to start immediately (a read-only transaction may stall), and the selected isolation level keeps row locking to a minimum. This could be explicitly stated by using the following command: $ RMU/EXTRACT- /TRANSACTION_TYPE=(WRITE,ISOLATION=READ_COMMITTED)- SAMPLE.RDB Using a transaction type of automatic adapts to different database settings: $ RMU/EXTRACT- /TRANSACTION_TYPE=(AUTOMATIC)- SAMPLE.RDB Example 18 This example shows the use of the Item=Workload qualifier to create a DCL command language script. $ RMU/EXTRACT/ITEM=WORKLOAD - SCRATCH/LOG/OPTION=(FILENAME,AUDIT) $! RMU/EXTRACT for Oracle Rdb V7.2-00 7-JAN-2006 22:00:42.72 $! $! WORKLOAD Procedure $! $!--------------------------------------------------------------------- $ SET VERIFY $ SET NOON $ $! Created on 7-JAN-2006 10:12:26.36 $! Last collected on 7-JAN-2006 22:00:34.47 $! $ RMU/INSERT OPTIMIZER_STATISTICS - SCRATCH - /TABLE=(CUSTOMERS) - /COLUMN_GROUP=(CUSTOMER_NAME) - /DUPLICITY_FACTOR=(4.0000000) - /NULL_FACTOR=(0.0000000) /LOG $ $! Created on 7-JAN-2006 10:12:26.36 $! Last collected on 7-JAN-2006 22:00:34.58 $! $ RMU/INSERT OPTIMIZER_STATISTICS - SCRATCH - /TABLE=(RDB$FIELDS) - /COLUMN_GROUP=(RDB$FIELD_NAME) - /DUPLICITY_FACTOR=(1.7794118) - /NULL_FACTOR=(0.0000000) /LOG $ . . . $ SET NOVERIFY $ EXIT Example 19 The following example shows the use of the Match option to select a subset of the workload entries based on the wildcard file name. $ RMU/EXTRACT/ITEM=WORKLOAD - SCRATCH/LOG/OPTION=(FILENAME,AUDIT,MATCH:RDB$FIELDS%) $! RMU/EXTRACT for Oracle Rdb V7.2-00 8-JAN-2006 10:53 $! $! WORKLOAD Procedure $! $!------------------------------------------------------------------------ $ SET VERIFY $ SET NOON $ ! Created on 7-JAN-2006 15:18:02.30 $ SET NOON $ $! Created on 7-JAN-2006 15:18:02.30 $! Last collected on 7-JAN-2006 18:25:04.27 $! $ RMU/INSERT OPTIMIZER_STATISTICS - SCRATCH - /TABLE=(RDB$FIELDS) - /COLUMN_GROUP=(RDB$FIELD_NAME) - /DUPLICITY_FACTOR=(1.0000000) - /NULL_FACTOR=(0.0000000) /LOG $ SET NOVERIFY $ EXIT Example 20 The following example shows the use of Item options Defer_ Constraints, Constraints, and Match to extract a table and its constraints. $ RMU/EXTRACT/ITEM=(TABLE,CONSTRAINT)- _$ /OPTION=(FILENAME_ONLY,NOHEADER,- _$ DEFER_CONSTRAINT,MATCH:EMPLOYEES%) - _$ MF_PERSONNEL set verify; set language ENGLISH; set default date format 'SQL92'; set quoting rules 'SQL92'; set date format DATE 001, TIME 001; attach 'filename MF_PERSONNEL'; create table EMPLOYEES ( EMPLOYEE_ID ID_DOM, LAST_NAME LAST_NAME_DOM, FIRST_NAME FIRST_NAME_DOM, MIDDLE_INITIAL MIDDLE_INITIAL_DOM, ADDRESS_DATA_1 ADDRESS_DATA_1_DOM, ADDRESS_DATA_2 ADDRESS_DATA_2_DOM, CITY CITY_DOM, STATE STATE_DOM, POSTAL_CODE POSTAL_CODE_DOM, SEX SEX_DOM, BIRTHDAY DATE_DOM, STATUS_CODE STATUS_CODE_DOM); comment on table EMPLOYEES is 'personal information about each employee'; alter table EMPLOYEES add constraint EMP_SEX_VALUES check(EMPLOYEES.SEX in ('M', 'F', '?')) deferrable add constraint EMP_STATUS_CODE_VALUES check(EMPLOYEES.STATUS_CODE in ('0', '1', '2', 'N')) deferrable alter column EMPLOYEE_ID constraint EMPLOYEES_PRIMARY_EMPLOYEE_ID primary key deferrable; commit work; Example 21 The following example shows the use of the option Group_Table to extract a table and its indexes: $ rmu/extract/item=(table,index)- _$ /option=(group_table,match=employees%,- _$ filename_only,noheader) db$:mf_personnel set verify; set language ENGLISH; set default date format 'SQL92'; set quoting rules 'SQL92'; set date format DATE 001, TIME 001; attach 'filename MF_PERSONNEL'; create table EMPLOYEES ( EMPLOYEE_ID ID_DOM constraint EMPLOYEES_PRIMARY_EMPLOYEE_ID primary key deferrable, LAST_NAME LAST_NAME_DOM, FIRST_NAME FIRST_NAME_DOM, MIDDLE_INITIAL MIDDLE_INITIAL_DOM, ADDRESS_DATA_1 ADDRESS_DATA_1_DOM, ADDRESS_DATA_2 ADDRESS_DATA_2_DOM, CITY CITY_DOM, STATE STATE_DOM, POSTAL_CODE POSTAL_CODE_DOM, SEX SEX_DOM, BIRTHDAY DATE_DOM, STATUS_CODE STATUS_CODE_DOM); comment on table EMPLOYEES is 'personal information about each employee'; create unique index EMPLOYEES_HASH on EMPLOYEES ( EMPLOYEE_ID) type is HASHED SCATTERED store using (EMPLOYEE_ID) in EMPIDS_LOW with limit of ('00200') in EMPIDS_MID with limit of ('00400') otherwise in EMPIDS_OVER; create unique index EMP_EMPLOYEE_ID on EMPLOYEES ( EMPLOYEE_ID asc) type is SORTED node size 430 disable compression; create index EMP_LAST_NAME on EMPLOYEES ( LAST_NAME asc) type is SORTED; commit work; alter table EMPLOYEES add constraint EMP_SEX_VALUES check(EMPLOYEES.SEX in ('M', 'F', '?')) deferrable add constraint EMP_STATUS_CODE_VALUES check(EMPLOYEES.STATUS_CODE in ('0', '1', '2', 'N')) deferrable; commit work; Example 22 The following example shows the output when you use the Item=Revoke_Entry qualifier: $ RMU/EXTRACT/ITEM=REVOKE_ENTRY ACCOUNTING_DB ... -- Protection Deletions -- -------------------------------------------------------------------------------- revoke entry on database alias RDB$DBHANDLE from [RDB,JAIN]; revoke entry on database alias RDB$DBHANDLE from [RDB,JONES]; revoke entry on database alias RDB$DBHANDLE from PUBLIC; revoke entry on table ACCOUNT from [RDB,JONES]; revoke entry on table ACCOUNT from PUBLIC; revoke entry on table ACCOUNT_BATCH_PROCESSING from [RDB,JONES]; revoke entry on table ACCOUNT_BATCH_PROCESSING from PUBLIC; revoke entry on table BILL from [RDB,JONES]; revoke entry on table BILL from PUBLIC; ... Example 23 The following example shows sample output for the WORK_STATUS table of MF_PERSONNEL. The uppercase DCL commands are generated by RMU Extract. $ RMU/EXTRACT/ITEM=UNLOAD- _$ /OPTION=(NOHEADER,FULL,MATCH:WORK_STATUS%) sql$database $ CREATE WORK_STATUS.COLUMNS ! Columns list for table WORK_STATUS ! in DISK1:[DATABASES]MF_PERSONNEL.RDB ! Created by RMU Extract for Oracle Rdb V7.2-00 on 1-JAN-2006 20:50:25.33 STATUS_CODE STATUS_NAME STATUS_TYPE $ RMU/UNLOAD - DISK1:[DATABASES]MF_PERSONNEL.RDB - /FIELDS="@WORK_STATUS.COLUMNS" - WORK_STATUS - WORK_STATUS.UNL $ $ EXIT $ RMU/EXTRACT/ITEM=LOAD- _$ /OPTION=(NOHEADER,FULL,MATCH:WORK_STATUS%) sql$database $ RMU/LOAD - /TRANSACTION_TYPE=EXCLUSIVE - /FIELDS="@WORK_STATUS.COLUMNS" - DISK1:[DATABASES]MF_PERSONNEL.RDB - WORK_STATUS - WORK_STATUS.UNL $ $ EXIT Example 24 The following example shows how to extract all constraints as an ALTER TABLE statement. $ rmu/extract/item=(notab,constr) db$:sql_personnel/opt=(nohead,mat=empl%,defer) set verify; set language ENGLISH; set default date format 'SQL92'; set quoting rules 'SQL92'; set date format DATE 001, TIME 001; attach 'filename $DISK1:[JONES]SQL_PERSONNEL.RDB'; alter table EMPLOYEES add constraint EMP_SEX_VALUES check((EMPLOYEES.SEX in ('M', 'F') or (EMPLOYEES.SEX is null))) initially deferred deferrable add constraint EMP_STATUS_CODE_VALUES check((EMPLOYEES.STATUS_CODE in ('0', '1', '2') or (EMPLOYEES.STATUS_CODE is null))) initially deferred deferrable alter column EMPLOYEE_ID constraint EMP_EMPLOYEE_ID_NOT_NULL not null initially deferred deferrable; 2 Insert_Optimizer_Statistics Inserts workload records into the RDB$WORKLOAD system relation. 3 Description When you enable and collect workload statistics, the system table RDB$WORKLOAD is created and populated. (See Collect_ Optimizer_Statistics for details.) You can update or delete these statistics using the RMU Collect Optimizer_Statistics command or the RMU Delete Optimizer_Statistics command, respectively. You might delete entries in the RDB$WORKLOAD table by accident or you might delete them to test how effective it is to maintain those particular workload statistics. If you decide that you want to maintain those deleted statistics, you can insert them with the RMU Insert Optimizer_Statistics command. To ensure that you insert accurate values, always issue an RMU Show Optimizer_ Statistics command with the Log qualifier before you issue an RMU Delete Optimizer_Statistics command. Refer to your generated log file for the values you should specify with the RMU Insert Optimizer_Statistics command. In addition you can use the RMU Insert Optimizer_Statistics command to create workload statistics in a copy of your master database. If you issue an RMU Collect Optimizer_Statistics command after having issued an RMU Insert Optimizer_Statistics command, statistics for the specified column groups are updated. 3 Format (B)0RMU Insert Optimizer_Statistics root-file-spec Command Qualifiers x Defaults x /Column_Group=(Column-list) x None - Required Qualifier /Duplicity_Factor=(floating-number) x /Duplicity_Factor=(1.0) /[No]Log[=file-spec] x See description /Null_Factor=(floating-number) x /Null_Factor=(0.0) /Tables=(table-list) x None - Required Qualifier 3 Parameters 4 root-file-spec root-file-spec Specifies the database into which optimizer statistics are to be inserted. The default file type is .rdb. 3 Command_Qualifiers 4 Column_Group Column_Group=(column-list) Specifies a list of columns that comprise a column group. You must use the Tables qualifier to specify the table or tables with which the columns are associated. The Column_Group=(column-list) qualifier is a required qualifier. 4 Duplicity_Factor Duplicity_Factor=(floating-number) Specifies the value to be inserted in the RDB$DUPLICITY_FACTOR column in the RDB$WORKLOAD table for the specified column group and table (or tables). The minimum value is 1.0 and the maximum value is the cardinality of the specified table. The default is the Duplicity_Factor=(1.0) qualifier. 4 Log Log Log=file-spec Nolog Specifies how the statistics inserted into the RDB$WORKLOAD system table are to be logged. Specify the Log qualifier to have the information displayed to SYS$OUTPUT. Specify the Log=file- spec qualifier to have the information written to a file. Specify the Nolog qualifier to prevent display of the information. If you do not specify any of variation of the Log qualifier, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 4 Null_Factor Null_Factor=floating-number Specifies the value to be inserted in the RDB$NULL_FACTOR column in the RDB$WORKLOAD table for the specified column group and table (or tables). The minimum value is 0.0 and the maximum value is 1.0. The default is the Null_Factor=(0.0) qualifier. 4 Tables Table Tables=(table-list) Specifies the table or tables for which column group entries are to be inserted. If you issue an RMU Collect Optimizer_Statistics command after you have inserted a workload column group into the RDB$WORKLOAD system table, those statistics are collected. The Tables=(table-list) qualifier is a required qualifier. 3 Usage_Notes o To use the RMU Insert Optimizer_Statistics command for a database, you must have the RMU$ANALYZE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o Cardinality statistics are automatically maintained by Oracle Rdb. Physical storage and workload statistics are only collected when you issue an RMU Collect Optimizer_Statistics command. To get information about the usage of physical storage and workload statistics for a given query, define the RDMS$DEBUG_FLAGS logical name to be "O". For example: $ DEFINE RDMS$DEBUG_FLAGS "O" When you execute a query, if workload and physical statistics have been used in optimizing the query, you will see a line such as the following in the command output: ~O: Workload and Physical statistics used o The Insert Optimizer_Statistics command modifies the RDB$LAST_ALTERED date of the RDB$WORKLOAD row so that it is activated for use by the optimizer. 3 Examples Example 1 The following example: 1. Collects workload statistics for the JOB_HISTORY table using the RMU Collect Optimizer_Statistics command 2. Deletes the statistics for one of the JOB_HISTORY workload column groups 3. Inserts the statistics that were just deleted into the RDB$WORKLOAD system table using the RMU Insert Optimizer_ Statistics command 4. Displays the current data stored in the RDB$WORKLOAD table for the JOB_HISTORY table using the RMU Show Optimizer_Statistics command $ RMU/COLLECT OPTIMIZER_STATISTICS MF_PERSONNEL.RDB - _$ /TABLE=(JOB_HISTORY)/STATISTICS=(WORKLOAD)/LOG Start loading tables... at 3-JUL-1996 10:54:04.16 Done loading tables.... at 3-JUL-1996 10:54:04.69 Start collecting workload stats... at 3-JUL-1996 10:54:06.76 Maximum memory required (bytes) = 6810 Done collecting workload stats.... at 3-JUL-1996 10:54:07.64 Start calculating stats... at 3-JUL-1996 10:54:07.84 Done calculating stats.... at 3-JUL-1996 10:54:07.86 Start writing stats... at 3-JUL-1996 10:54:09.34 --------------------------------------------------------------------- Optimizer Statistics collected for table : JOB_HISTORY Workload Column group : EMPLOYEE_ID Duplicity factor : 2.7400000 Null factor : 0.0000000 Workload Column group : EMPLOYEE_ID, JOB_CODE, JOB_START, JOB_END, DEPARTMENT_CODE, SUPERVISOR_ID Duplicity factor : 1.5930233 Null factor : 0.3649635 Done writing stats.... at 3-JUL-1996 10:54:09.90 $ RMU/DELETE OPTIMIZER_STATISTICS MF_PERSONNEL.RDB - _$ /TABLE=(JOB_HISTORY)/COLUMN_GROUP=(EMPLOYEE_ID,JOB_CODE, - _$ JOB_START,JOB_END,DEPARTMENT_CODE,SUPERVISOR_ID)/LOG Changing RDB$SYSTEM area to READ_WRITE. Workload column group deleted for JOB_HISTORY : EMPLOYEE_ID, JOB_CODE, JOB_START, JOB_END, DEPARTMENT_CODE, SUPERVISOR_ID $ ! $ RMU/INSERT OPTIMIZER_STATISTICS MF_PERSONNEL.RDB - _$ /TABLE=(JOB_HISTORY) /COLUMN_GROUP=(EMPLOYEE_ID,JOB_CODE, - _$ JOB_START,JOB_END,DEPARTMENT_CODE,SUPERVISOR_ID) - _$ /DUPLICITY_FACTOR=(1.5930233)/NULL_FACTOR=(0.3649635)/LOG Changing RDB$SYSTEM area to READ_WRITE. Workload column group inserted for JOB_HISTORY : EMPLOYEE_ID, JOB_CODE, JOB_START, JOB_END, DEPARTMENT_CODE, SUPERVISOR_ID $ ! $ RMU/SHOW OPTIMIZER_STATISTICS MF_PERSONNEL.RDB - _$ /TABLE=(JOB_HISTORY)/STATISTICS=(WORKLOAD)/LOG -------------------------------------------------------------------- Optimizer Statistics for table : JOB_HISTORY Workload Column group : EMPLOYEE_ID Duplicity factor : 2.7400000 Null factor : 0.0000000 First created time : 3-JUL-1996 10:37:36.43 Last collected time : 3-JUL-1996 10:54:09.62 Workload Column group : EMPLOYEE_ID, JOB_CODE, JOB_START, JOB_END, DEPARTMENT_CODE, SUPERVISOR_ID Duplicity factor : 1.5930233 Null factor : 0.3649635 First created time : 3-JUL-1996 10:57:47.65 Last collected time : 3-JUL-1996 10:57:47.65 2 Librarian Allows you to list or remove backed up Oracle Rdb databases from a Librarian utility. 3 Description The RMU Librarian command allows you to list or remove backed up Oracle Rdb databases from a Librarian utility that conforms to the Oracle Media Manager interface. You cannot perform both the list and remove operations within one command. 3 Format (B)0RMU/Librarian backup-file-spec Command Qualifiers x Defaults x /List[=(Output=file-name),(options,...)] x See Description /Remove[=[No]Confirm,(options,...)] x See Description 3 Parameters 4 backup-file-spec Identifies the backed up Oracle Rdb database previously stored in the Librarian utility. Use the same backup file name that was used in the Oracle RMU Backup command. A default file type of .RBF is assumed if none is specified. Any device, directory, or version number specified with the backup file name will be ignored. If the Librarian utility supports wild card characters, you can use them for the backup file name when you use the List qualifier. Wild card characters cannot be used with the Remove qualifier. 3 Command_Qualifiers 4 List List List=Output=file-name List=options Allows you to display a backed up Oracle Rdb database stored in a Librarian utility. If you use the List qualifier without the Output option, the output is sent to the default output device. If you use the Output option, the output is sent to the specified file. All data streams existing in the Librarian that were generated for the specified backup name will be listed. The information listed for each data stream can include: o The backup stream name based on the backup file. o Any comment associated with the backup stream name. o The creation method associated with the backup stream name. This will always be STREAM. o The creation data and time when the stream was backed up to the Librarian. o Any expiration date and time specified for deletion of the stream by the Librarian. o The media sharing mode that indicates if the media can be accessed concurrently or not. o The file ordering mode that indicates if files on the media can be accessed in random order or sequential order. o Any volume labels for the media that contain the backup stream. Not all of these items will be listed depending on the particular Librarian utility. The List qualifier can accept the following options: o Trace_File=file-specification The Librarian application writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian application. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian application. The pre-determined values are: - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical-name=equivalence-value,...) Use this option to specify a list of process logical names that the Librarian application can use to specify catalogs or archives for listing or removing backup files, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of the logical names. The list of process logical names is defined by Oracle RMU prior to the start of the list or remove operation. 4 Remove Remove Remove=Confirm Remove=Noconfirm Remove=options Allows you to delete all data streams existing in the Librarian that were generated for the specified backup file. This command should be used with caution. You should be sure that a more recent backup for the database exists in the Librarian under another name before you use this command. The Confirm option is the default. It prompts you to confirm that you want to delete the backup from the Librarian. If you do not want to be prompted, use the Noconfirm option. The deletion will be performed with no confirmation prompt. The Remove qualifier can accept the following options: o Trace_File=file-specification The Librarian application writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian application. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian application. The pre-determined values are: - Level 0, trace all error conditions, is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical-name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian application can use to specify catalogs or archives for listing or removing backup files, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of the list or remove operation. 2 Load There are two RMU Load commands, as follows: o An RMU Load command without the Plan qualifier allows you to load data into the database you specify as a parameter to the Load command. o An RMU Load command with the Plan qualifier allows you to execute a plan file you specify as a parameter to the Load command. 3 Database Loads data into the tables of the database. You can use the RMU Load command to: o Perform the initial load of an Oracle Rdb database. o Reload a table after performing a restructuring operation. o Load an archival database. o Move data from one database to another. o Load security audit records from an OpenVMS security audit table into the database being audited, or into a different database than the one being audited. o Load additional rows into an existing table. (However, note that it cannot be used to modify existing rows.) o Import data into a database from an application that generates RMS files. You can load data using either of the following two methods: o A single-process method This was the only method available prior to Oracle Rdb V7.0. The single process method uses one process to both read the input file and load the target table. o A multiprocess method, also called a parallel load The parallel load method, which you specify with the Parallel qualifier, enables Oracle RMU to use your process to read the input file and use one or more executors (subprocesses or detached slave process, depending on additional factors) to load the data into the target table. This results in concurrent read and write operations, and in many cases, substantially improves the performance of the load operation. By default, Oracle RMU sets up a parallel load operation as follows: o Your process serves as the load operation execution manager. o Each storage area (partition) in the table being loaded is assigned an executor. o Each executor is assigned four communications buffers. (You can override this default with the Buffer_Count option to the Parallel qualifier.) o Each communications buffer holds the number of rows defined by the Row_Count qualifier. Once the executors and communications buffers are set up, the parallel load operation processes the input file as follows: 1. Your process begins reading the input file and determines the target storage area for each row in the input file. 2. Your process places each row in the communications buffer for the executor assigned to the data's target storage area. 3. When an executor's first communications buffer becomes full, it begins loading the data into the target storage area. 4. If your process has another portion of data ready for a given executor before that executor has completed loading its first buffer of data, your process places the next portion of data in the second communications buffer for that executor. 5. Each executor, concurrent with each of the other executors, loads the data from its buffers. 6. Your process continues reading, sorting, and assigning data to each executor (by placing it in that executor's communication buffer) until all the data from the input file has been sorted, assigned, and loaded. The Row_Count qualifier and Parallel qualifier (which provides the Executor_Count and Buffer_Count options) give you the ability to fine tune the Parallel load operation. See the Oracle Rdb Guide to Database Design and Definition for tips on optimizing the performance of the load operation. 4 Description The RMU Load command accepts the following five types of data files, all of which, except the security audit journal, have the file extension .unl: o Text data file o Delimited text data file o Binary data file o Specially structured file o OpenVMS security audit journal file With the exception of the specially structured file and the security audit journal file, you must provide a record definition file (.rrd) on the RMU Load command line to load these data files. The record definition file provides Oracle RMU with a description of (metadata for) the data you are loading. The following list describes the additional requirements for loading each of these types of files: o Text data file To load a text data file (.unl), you must specify the Record_ Definition qualifier with the Format=Text option. The following command loads text data (employees.unl) into the EMPLOYEES table of the mf_personnel database. The employees.rrd file provides the record definition for the data in employees.unl $ RMU/LOAD/RECORD_DEFINITION=(FILE=employees.rrd, FORMAT=TEXT) - _$ mf_personnel EMPLOYEES employees.unl You can generate an appropriate .rrd file for the preceding example by issuing the following command: $ RMU/UNLOAD/RECORD_DEFINITION=(FILE=employees.rrd, FORMAT=TEXT) - _$ mf_personnel EMPLOYEES unload.unl o Delimited text data files To load delimited text data files (.unl) you must specify the Record_Definition qualifier with the with the Format=Delimited_Text option. The following command loads delimited text data (employees.unl) into the EMPLOYEES table of the mf_personnel database. The employees.rrd file describes the format of employees.unl $ RMU/LOAD/RECORD_DEFINITION=(FILE=employees.rrd, - _$ FORMAT=DELIMITED_TEXT, TERMINATOR="#") - _$ mf_personnel EMPLOYEES employees.unl You can generate an appropriate .rrd file for the preceding example by issuing the following command: $ RMU/UNLOAD/RECORD_DEFINITION=(FILE=employees.rrd, - _$ FORMAT=DELIMITED_TEXT) mf_personnel EMPLOYEES unload.unl o Binary data files To load binary data files, you must ensure that the records you load match the record definition in both size and data type. The records must all have the same length and the data in each record must fill the entire record. If the last field is character data and the information is shorter than the field length, the remainder of the field must be filled with spaces. You cannot load a field that contains data stored in packed decimal format. The following command loads binary data (employees.unl) into the EMPLOYEES table of the mf_personnel database. The employees.rrd file describes the format of employees.unl. $ RMU/LOAD/RECORD_DEFINITION=(FILE=employees.rrd) mf_personnel - _$ EMPLOYEES employees.unl You can generate an appropriate .rrd file for the preceding example by issuing the following command: $ RMU/UNLOAD/RECORD_DEFINITION=(FILE=employees.rrd) mf_personnel - _$ EMPLOYEES unload.unl o Specially structured binary files that include both data and metadata. To load the specially structured binary files (created by the RMU Unload command without the Record_Definition qualifier) you must specify the file (.unl) created by the RMU Unload command. The following command loads the binary data contained in the employees.unl file into the EMPLOYEES table of the mf_ personnel database. The record definition information is contained within the binary .unl file. $ RMU/LOAD MF_PERSONNEL EMPLOYEES employees.unl This specially structured employees.unl file is created with the following RMU Unload command: $ RMU/UNLOAD MF_PERSONNEL EMPLOYEES employees.unl o Security audit journal files To load the records from a security audit journal file maintained by the OpenVMS operating system, you must decide whether to load records into the same database for which security audit journal records are being recorded or to load them into a separate database. In either case you do not need to specify a record definition file; use of the Audit qualifier indicates to Oracle RMU that the record definition is that of the security audit journal file. The following command loads the records from the security audit journal file (with a logical name of SECURITY_AUDIT) for the mf_personnel database into the AUDIT_TABLE table of the mf_personnel database: $ RMU/LOAD/AUDIT MF_PERSONNEL.RDB AUDIT_TABLE - _$ SECURITY_AUDIT This example loads the records from the security audit journal file (with a logical name of SECURITY_AUDIT) for the mf_ personnel database into the AUDIT_TABLE table of the audit database: $ RMU/LOAD/AUDIT=DATABASE_FILE=MF_PERSONNEL.RDB AUDIT.RDB - _$ AUDIT_TABLE SECURITY_AUDIT See the Usage Notes for more detailed information on loading security audit journal records and the file name of the security audit journal. In all cases where you specify a record definition file (.rrd), the record definition file and the database definition of the table being loaded must match in the number of specified fields and the data type of each field. If the data you want to load has more fields than the database table definition specifies, you can still load the data, but you must use the FILLER keyword with the field definition in your .rrd file to represent the additional field. See Example 15 in the Examples help entry under this command. By default, the table specified in the RMU Load command is reserved for PROTECTED WRITE. Data Type Conversions Performed by Oracle Rdb shows the data type conversions that can occur while you are performing a load or unload operation. Table 11 Data Type Conversions Performed by Oracle Rdb Original Data Type New Data Type TINYINT INTEGER, QUADWORD, SMALLINT, FLOAT, DOUBLE PRECISION, VARCHAR, CHAR SMALLINT INTEGER, QUADWORD, FLOAT, DOUBLE PRECISION, VARCHAR, CHAR INTEGER SMALLINT, QUADWORD, FLOAT, DOUBLE PRECISION, VARCHAR, CHAR QUADWORD SMALLINT, INTEGER, FLOAT, DOUBLE PRECISION, VARCHAR, CHAR FLOAT DOUBLE PRECISION, CHAR, and VARCHAR DOUBLE FLOAT, CHAR, and VARCHAR PRECISION DATE CHAR or VARCHAR TIME CHAR or VARCHAR TIMESTAMP CHAR or VARCHAR INTERVAL CHAR or VARCHAR CHAR FLOAT, DOUBLE PRECISION, DATE, TIME, TIMESTAMP, INTERVAL, VARCHAR, SMALLINT, INTEGER, or QUADWORD See the Oracle Rdb SQL Reference Manual for a description of these data types. 4 Format (B)0 RMU/Load root-file-spec table-name input-file-name Command Qualifiers x Defaults x /Audit[=Database_File=db-name] x No audit table loaded /Buffers=n x See description /Commit_Every=n x See description /[No]Constraints[=Deferred] x /Constraints /Corresponding x See description /[No]Defer_Index_Updates x /Nodefer_Index_Updates /[No]Dialect=(dialect-opts) x /Dialect=SQL99 /[No]Execute x /Execute /Fields=(column-name-list) x See description /List_Plan=output-file x See description /[No]Log_Commits x /Nolog_Commits /[No]Match_Name=table-name x /Nomatch_Name /Parallel[=(options)] x See description /[No]Place x /Noplace /Record_Definition= x See description ({File|Path}=name[,options]) x (B)0 /[No]Restricted_Access x /Norestricted_Access /Row_Count=n x See description /[No]Skip=n x /Noskip /Statistics=(stat-opts) x See description /Transaction_Type=Share-mode x Protected /[No]Trigger_Relations[=(table_name_list)] x /Trigger_Relations /[No]Virtual_Fields[=[No]Automatic] x /Novirtual_Fields 4 Parameters 5 root-file-spec The file specification for the database root file into which the table will be loaded. The default file extension is .rdb. 5 table-name The name of the table to be loaded, or its synonym. When the Audit qualifier is specified, the table-name parameter is the name of the table in which you want the security audit journal records to be loaded. If the table does not exist, the RMU Load command with the Audit qualifier creates the table and loads it. If the table does exist, the RMU Load command with the Audit qualifier loads the table. 5 input-file-name The name of the file containing the data to be loaded. The default file extension is .unl. When the Audit qualifier is specified, the input-file-name parameter is the name of the journal containing the audit record data to be loaded. The default file extension is .AUDIT$JOURNAL. You can determine the name of the security audit journal by using the DCL SHOW AUDIT/JOURNAL command. 4 Command_Qualifiers 5 Audit Audit Audit=Database_File=db-name Allows you to load a database's security audit records from an OpenVMS security audit journal into one of the following: o A table in the database being audited Specify the Audit qualifier without the Database_File option to indicate that you want the security audit records to be loaded into the database specified with the root-file-spec parameter. o A table in a different database than the one being audited Specify the Audit=Database_File=db-name qualifier to indicate that you want to security audit records for the database specified with the root-file-spec command parameter to be loaded into the database specified with the db-name option parameter. If you specify the Audit qualifier, you cannot specify the Fields or Trigger_Relations qualifiers. In addition you cannot specify the Audit qualifier with a parallel load operation. If you attempt to do so, Oracle RMU issues a warning and performs a single-executor load operation. 5 Buffers Buffers=n Specifies the number of database buffers used for storing data during the load operation. If no value is specified, the default value for the database is used. (The default value for the database is defined by the logical name RDM$BIND_BUFFERS, or if the logical is not defined, can be determined by using the RMU Dump command with the Header qualifier. The RDM$BIND_BUFFERS logical name, if defined, overrides the value displayed with the RMU Dump command.) Fewer I/O operations are required if you can store as much data as possible in memory when many indexes or constraints are defined on the target table. Therefore, specify more buffers than allowed by the default value to increase the speed of the load operation. See the Oracle Rdb7 Guide to Database Performance and Tuning for detailed recommendations on setting the number of database buffers. 5 Commit_Every Commit_Every=n Specifies the frequency with which Oracle Rdb commits the data being loaded. For a single-executor load operation, Oracle Rdb commits the data after every n records that are stored. The default is to commit only after all records have been stored. For a parallel load operation, the Commit_Every qualifier applies separately to each of the executors (processes) used. For example, if five parallel processes are running, and the Commit_Every=2 qualifier is specified, Oracle RMU commits data for each process after it has stored 2 records. This means that if the Commit_Every=1000 qualifier is specified when you load one million records with 10 parallel processes, the .ruj files will store up to 10,000 rows of before-image data. If you specify the Defer_Index_Updates qualifier and a high value for the Commit_Every qualifier, memory requirements are high. See the description of the Defer_Index_Updates qualifier for details. Commit operations may occur more frequently than you specify under certain conditions. See the description of the Defer_Index_ Updates qualifier for details. To determine how frequently you should commit data, decide how many records you are willing to reload if the original load operation fails. If you use the Statistics=On_Commit qualifier, you receive a message indicating the number of records loaded at each commit operation. Then, if a failure occurs, you know where to resume loading. If you specify the Place qualifier and a failure occurs, resume loading at the point of the previous commit, instead of the record number of the last successful commit. The Place qualifier restructures the .unl file prior to loading, so the record number on which the load operation failed does not correspond to the same number in the original .unl file. 5 Constraints Constraints Constraints=Deferred Noconstraints Specifies when or if constraints are evaluated for data being loaded. If you specify the Constraints qualifier, constraints are evaluated as each record is loaded. If you specify the Noconstraints qualifier, constraints are not evaluated at all during the load operation. If you specify the Constraints=Deferred qualifier, constraints are evaluated after all data from the input file has been loaded. The default is the Constraints qualifier. Oracle Corporation recommends that you accept the default for most load operations. The Noconstraints and Constraints=Deferred qualifiers are useful when load performance is your highest priority, you fully understand the constraints defined for your database, and you are familiar enough with the input data to be fairly certain that loading that data will not violate constraints; then you might use these qualifiers as follows: o Constraints=Deferred This qualifier is particularly useful for improving performance when you are loading data into a new table. Oracle Corporation strongly recommends that you issue an RMU Verify command with the Constraints qualifier when the load operation has completed. Note, however, that issuing the RMU Verify command after the load operation has completed takes about the same amount of time that would have been spent had you specified the RMU Load command with the Constraints qualifier. In other words, by specifying the Constraints=Deferred qualifier, you are only delaying when the constraint verification will take place. o Noconstraints This qualifier is particularly useful when you are performing a parallel load operation with the Defer_Index_Updates qualifier. Oracle Corporation strongly recommends that you issue an RMU Verify command with the Constraints qualifier when the load operation has completed. Note, however, that when you issue the RMU Verify command with the Constraints qualifier, all rows in the table are checked for constraint violations, not just the rows that are loaded. Consider the following before issuing an RMU Load command with the Noconstraints or Constraints=Deferred qualifier: o If a table is populated with data prior to a load operation, it is less expensive to check constraints on each record as it is being loaded, than to verify constraints on the entire table after the set of new records has been loaded. For example, assume you load 200 new records into a table that currently holds 2,000 records and one constraint is defined on the table. If you verify constraints as the records are being loaded, constraint validation is performed 200 times. If you wait and verify constraints after the load operation completes, constraint verification must be performed for 2,200 records o If an RMU Verify command reveals that constraint violations occurred during the load operation, you must track down those records and either remove them or make other modifications to the database to restore the data integrity. This can be a time-consuming process. Also consider a situation where all of the following are true: o You perform a parallel load operation o You specify the Constraints qualifier o The table into which you are loading data has a constraint defined on it o The constraint defined on the table was defined as deferred o Constraint evaluation fails during the load operation In a case such as the preceding, you can not easily determine which rows were loaded and which were not. Therefore Oracle Corporation recommends that if deferred constraints are defined on a table, then you should also specify the Constraints=Deferred qualifier in your parallel load command. When you follow this recommendation, the records that violate the constraint are stored in the database. When the load operation completes, you can remove from the database those records that violate the constraint. See Example 6 in Verify for an example of the steps to take if an RMU Verify command reveals that an RMU Load command has stored data that violates constraints into your database. 5 Corresponding Corresponding Loads fields into a table from the .unl file by matching the field names in the .rrd file to the column names in the table. The Corresponding qualifier makes it more convenient to unload, restructure, and reload a table. For example, if the columns in the table appear in the order: EMPLOYEE_ID, LAST_NAME, FIRST_NAME, but the data in your .unl file appears in the order: EMPLOYEE_ID, FIRST_NAME, LAST_NAME, and your .rrd file lists the fields in the order: EMPLOYEE_ID, FIRST_NAME, LAST_NAME, you can use the Corresponding qualifier to load the data in your .unl file correctly. (You could also use the Fields qualifier to accomplish the same task, but this can get tedious if there are numerous fields.) The .unl file must contain data for each field in the database into which it is being loaded; if it does not, you should use the Fields qualifier. If the Corresponding qualifier is omitted, the RMU Load command loads the data into database fields by the ordinal position in which they appear in the .unl, not by the column name described in the .rrd file. The Corresponding qualifier cannot be used with either the Fields or Audit qualifiers. 5 Defer_Index_Updates Defer_Index_Updates Nodefer_Index_Updates The Defer_Index_Updates qualifier specifies that non-unique indexes (other than those that define the placement information for data in a storage area) will not be rebuilt until commit time. Use of this qualifier results in less I/O and fewer lock conflicts than when index builds are not deferred, but results in a total failure of a load operation if any lock conflicts are encountered. In such a case, the entire load operation is rolled back to the previous commit and you must repeat the load operation. (Record insertion recommences at the beginning of the input file). For this reason, you should only use the Defer_ Index_Updates qualifier when all of the following are true: o You specify the Noconstraints qualifier (or you have dropped constraints, or no constraints are defined on the table). o You have dropped triggers from the table (or triggers are not defined for the table). o No other users are accessing the table being loaded. Also be aware that required virtual memory can be quite large when you defer index updates. Required virtual memory is directly proportional to the following: o The length of the Ikeys in the indexes being deferred o The number of indexes being deferred o The value for n specified with the Commit_Every qualifier You can estimate the amount of virtual memory required for each deferred index using the following formula, where: o n = the value specified with the Commit_Every qualifier o I = (length of the Ikey + 50) n * (I * number_defered_ikeys) The Nodefer_Index_Updates qualifier is the default. When you specify the Nodefer_Index_Updates qualifier (or accept the default), both the indexes that define the placement information for data in a storage area and any other indexes defined on the table being loaded are rebuilt at verb time. This can result in a managed deadlock situation when the Parallel qualifier is specified. The following describes such a scenario: o Executor_1 locks index node A in exclusive mode o Executor_2 locks index node B in exclusive mode o Executor_1 requests a lock on index node B o Executor_2 requests a lock on index node A In such a situation, Oracle Rdb resolves the deadlock by directing one of the executors to commit the data it has already stored. This resolves the deadlock situation and the load operation continues. 5 Dialect Dialect Nodialect The Dialect qualifier is used to control whether truncation of string data during the loading of data is reported or not. This loss of data might be significant. RMU Load defaults to SQL dialect SQL99 which implicitly checks for and reports truncations during INSERT operations. o /NODIALECT, /DIALECT=SQL89 or /DIALECT=NONE will not report any truncation errors, which is the "old" behavior of Rdb (prior to July 2008). o /DIALECT=SQL99 (the default) will enable reporting of truncation errors. Note that truncation occurs if non-space characters are discarded during the insert. 5 Execute Execute Noexecute The Execute and Noexecute qualifiers are used with the List_Plan qualifier to specify whether or not the generated plan file is to be executed. The Noexecute qualifier specifies that the plan file should be created but should not be executed. Regardless of whether you use the Noexecute or Execute qualifier (or accept the default), Oracle RMU performs a validity check on the RMU Load command you specify. The validity check determines such things as whether the specified table is in the specified database, the .rrd file (if specified) matches the table, and that the number of columns specified with the Fields qualifier matches the number of columns in the .unl file. The validity check does not determine such things as whether your process and global page quotas are sufficient. By default, the plan file is executed when an RMU Load command with the List_Plan qualifier is issued. 5 Fields Fields=(column-name-list) Specifies the column or columns of the table to be loaded into the database. If you list multiple columns, separate the column names with a comma, and enclose the list of column names within parentheses. Also, this qualifier specifies the order of the columns to be loaded if that order differs from the order defined for the table. The number and data type of the columns specified must agree with the number and data type of the columns in the .unl file. The default is all columns defined for the table in the order defined. If you specify an options file in place of a list of columns, and the options file is empty, the RMU Load command loads all fields. 5 List_Plan List_Plan[=output-file] Specifies that Oracle RMU should generate a plan file and write it to the specified output file. A plan file is a text file that contains all the qualifiers specified on the RMU Load command line. In addition, it specifies the executor names (if you are performing a parallel load operation), the directory for the .ruj files, the exception files, and the file created by the Place_ Only qualifier (if specified). Oracle RMU validates the Oracle RMU command prior to generating the plan file to ensure that an invalid plan file is not created. (This is true regardless of whether or not you specify the Noexecute qualifier.) For example, the following command is invalid and returns an error message because it specifies conflicting qualifiers (Corresponding and Fields): $ RMU/LOAD/RECORD_DEF=FILE=NAMES.RRD/CORRESPONDING - _$ /FIELDS=(LAST_NAME, FIRST_NAME)/LIST_PLAN=my_plan.plan - _$ MF_PERSONNEL.RDB EMPLOYEES NAMES.UNL %RMU-F-CONFLSWIT, conflicting options CORRESPONDING and FIELDS... See the description of the Execute qualifier for a description of what items are included when Oracle RMU validates the RMU Load command. See the Examples section for a complete example and description of a plan file. You can use the generated plan as a starting point for building a load operation that is tuned for your particular configuration. The output file can be customized and then used with subsequent load operations as the parameter to the RMU Load Plan command. See Load Plan for details. If you want to create only a load plan file and do not want to execute the load plan when the RMU Load command is issued, specify the Noexecute qualifier. When you specify the Noexecute qualifier, you must specify a valid Oracle RMU command. One way to prototype a plan file prior to creating a potentially very large .unl file is to specify the List_Plan qualifier and the Noexecute qualifier along with a valid record definition (.rrd) file and an empty .unl file on the RMU Load command line. The .rrd file contains the information Oracle RMU needs to perform the validation of the plan file; however, because data is not loaded when you specify the Noexecute qualifier, Oracle RMU does not attempt to load the .unl file. Note, however, that you cannot specify the Fields qualifier when using this strategy. (When you specify the Fields qualifier, Oracle RMU checks to make sure the number of columns specified with the Fields qualifier match the number of columns specified in the .unl file.) If you do not specify a file extension, the default file extension for the plan file is .plan. 5 Log_Commits Log_Commits Nolog_Commits Causes a message to be printed after each commit operation. In the case of a parallel load, a message is printed after each executor commits. The default is the Nolog_Commits qualifier, where no message is printed after individual commit operations. The Nolog_Commits qualifier does, however, cause a commit operation total to be printed after the operation completes or generates an error. 5 Match_Name Match_Name=table-name Nomatch_Name Specifies the table name to be read. Tables exported by SQL into an interchange file can be individually loaded into a database. The default behavior of the RMU Load command is to locate and load the first set of table data in the unload file. If this is not the table you want, you can use the Match_Name qualifier to specify a different table name. If the Match_Name qualifier is used without a table-name, Oracle RMU assumes the name of the table being loaded is also the name of the table in the source data file. The default is the Nomatch_Name qualifier. 5 Parallel Parallel[=(options)] Specifies a parallel load operation. A parallel load operation is especially effective when you have large partitioned tables that do not contain segmented strings and for which no constraints or triggers are defined. If you specify the Parallel qualifier without any options, your load operation is assigned one executor and four communications buffers for that executor. A communications buffer is used for communications between your process and the executors. If you want to assign additional executors or communications buffers, or both, use one or both of the following options: o Buffer_Count=n Allows you to specify the number of communications buffers assigned to each executor in a parallel load operation. Do not confuse this with the Buffers=n qualifier. The Buffers=n qualifier specifies the number of database buffers to use during the load operation. o Executor_Count=n Allows you to specify the number of worker processes to be assigned to the load operation. Ideally, the number of executors should be equal to the number of table partitions. You should not assign a greater number of executors than the number of table partitions. If a table is randomly or vertically partitioned, Oracle RMU creates only one executor, regardless of the number you specify. If the user account's MAXDETACH UAF value is greater than 0, then executors are created as detached processes. If there is no MAXDETACH value set, then executors are created as subprocesses. (A MAXDETACH value = 0 equates to unlimited detached processes.) At the end of each load operation, Oracle RMU displays summary statistics for each executor in the load operation and the main process. Look at the "Idle time" listed in the statistics at the end of the job to detect data skew and look at "Early commits" to detect locking contention. If some executors have a large amount of idle time, you likely have data that is skewed. Ideally, data loaded with the Parallel qualifier should appear in random order within the .unl file. Data that is already in partition order when you attempt to perform a parallel load operation results in high idle time for each executor and thus defeats the advantages of a parallel load operation. The summary statistics also list the number of records read from the input file, the number of data records stored, and the number of data records rejected. In most cases, the number of data records rejected plus the number of data records stored equals the number of data read from the input file. However, under the following circumstances this equation does not hold: o The parallel load operation aborts due to a duplicate record that is not allowed. o You did not specify an exception file. Similarly if a load operation aborts due to a record in the input file being improperly delimited for a delimited text load, the records rejected plus the records stored do not equal the number of records read from the input file. You cannot use a parallel load operation to load list data (segmented string) records or security audit records. If you specify a parallel load operation and attempt to load list data or security audit records, Oracle RMU returns a warning and performs a single-process (non-parallel) load operation. 5 Place Place Noplace Sorts records by target page number before they are stored. The Place qualifier automatically builds an ordered set of database keys (dbkeys) when loading data and automatically stores the records in dbkey order, sequentially, page by page. During a parallel load operation, each worker executor builds its own ordered set of dbkeys. The number of work files used by the RMU Load command is controlled by the RDMS$BIND_SORT_WORKFILES logical name. The allowable values are 1 through 10 inclusive, with a default value of 2. The location of these work files can be specified with device specifications, using the SORTWORKn logical name (where n is a number from 0 to 9). See the OpenVMS documentation set for more information on using SORT/MERGE. See the Oracle Rdb7 Guide to Database Performance and Tuning for more information on using these Oracle Rdb logical names. A significant performance improvement occurs when the records are stored by means of a hashed index. By using the Commit_ Every qualifier with the Place qualifier, you can specify how many records to load between COMMIT statements. Performance may actually decrease when records are stored by means of a sorted index. The default is the Noplace qualifier. 5 Record_Definition Record_Definition=(File=name[,options]) Record_Definition=(Path=name[,options]) Specifies the RMS record definition or the data dictionary record definition to be used when data is loaded into the database. Use the File=name parameter to specify an RMS record definition file; use the Path=name parameter to specify that the record definition be extracted from the data dictionary. (If the record definition in the data dictionary contains variants, Oracle RMU will not be able to extract it.) The default file extension for the File=name parameter is .rrd. The syntax for the .rrd file is similar to that used by the Common Dictionary Operator (CDO) interface for the data dictionary. You must define columns before you can define rows. You can place only one column on a line. You can create a sample .rrd file by using the RMU Unload command with the Record_ Definition qualifier. You must ensure that the record definition in the .rrd file and the actual data are consistent with each other. Oracle Rdb does not check to see that data types in the record definition and the data match. See the help entry for RRD_File_Syntax and the Oracle Rdb Guide to Database Design and Definition for more information about the format of the .rrd file. You must specify either the File=name or Path=name parameter. The options available are: o Exception_File=exception-file Allows you to write unloadable records to a single exception file for a single-process load operation and into multiple exception files for a parallel load operation. If you generate a load plan for a parallel load operation, each executor is assigned its own exception file. In this case, the exception- file name you specify is given a different file extension for each executor. While Oracle RMU is loading data from an RMS file, if an exception file is specified, then under certain circumstances an invalid record in the input file does not cause the RMU Load command to abort. Instead, Oracle RMU creates the exception file (or files), writes the unloadable record into this exception file (or files), and continues loading the remaining records. This process occurs only if the data is invalid on the actual insert, due to index, constraint, or trigger errors. If the record has an invalid format in the RMS file (for example, a missing delimiter), the exception file is not used, and the load process aborts. At the end of the load operation, you can process the exception file (or files) to correct any problems, and then reload directly from the exception file or files. The load operation gives an informational message for each of the unloadable records and also gives a summary of the number of records stored and the number of records rejected. All records that could not be loaded will be written into the file or files as specified with the argument to the Exception_ File option. The default file extension for the exception file is .unl for single-process loads; for parallel loads the default extension is EXC_n, where n corresponds to the executor number assigned by Oracle RMU. The exception file or files are created only if there are unloadable records. If the Exception_File option is not specified, no exception files are created, and the load operation aborts at the first occurrence of an exception. However, note that if the Defer_Index_Updates qualifier is specified, and a constraint violation or lock conflict occurs, the load operation aborts when it attempts to commit the transaction. If the Defer_Index_Updates qualifier is not specified, records that cause a constraint violation are written to the exception file or files and the load operation continues loading the remaining records. o Format=Text If you specify the Format=Text option, Oracle RMU converts all data to printable text before loading it. o If you do not specify the Format option, then Oracle RMU expects to load a fixed-length binary flat file. The data type of the fields must be specified in the .rrd file. o Format=(Delimited_Text [,delimiter-options]) If you specify the Format=Delimited_Text option, the .rrd file contains only text fields and specifies the maximum length of the columns in the file containing delimited ASCII text. The column values that are longer than those specified in the .rrd file are truncated. Note that DATE VMS types must be specified in the collatable time format, which is yyyymmddhhmmsscc. For example, March 20, 1993 must be specified as: 1993032000000000. Unless you specify the Format=Delimited_Text option, delimiters are regarded as part of the data by Oracle RMU. Example 13 in the Examples help entry under this command demonstrates the Format=Delimited_Text option. Delimiter options (and their default values if you do not specify delimiter options) are as follows. Note that with the exception of the Prefix and Suffix delimiter options, the values specified must be unique. The Prefix and Suffix values can be the same value as each other, but not the same as other delimiter options. The Null string must also be unique. - Prefix=string Specifies a prefix string that begins any column value in the ASCII input file. If you omit this option, the column prefix is assumed to consist of a quotation mark ("). - Separator=string Specifies a string that separates column values of a row. If you omit this option, the column separator is assumed to consist of a single comma (,). - Suffix=string Specifies a suffix string that ends any column value in the ASCII input file. If you omit this option, the column suffix is assumed to consist of a quotation mark ("). - Terminator=string Specifies the row terminator that completes all the column values corresponding to a row. If you omit this option, the row terminator is assumed to be the end of the line. - Null=string Specifies a string, which when found in the input record, is stored as NULL in the database column. This option is only valid when the Delimited_Text option is specified also. The Null option can be specified on the command line as any one of the following: * A quoted string * An empty set of double quotes ("") * No string If provided, the string that represents the null character must be quoted on the Oracle RMU command line, however, it must not be quoted in the input file. You cannot specify a blank space or spaces as the null character. If the final column or columns of a record are to be set to NULL, you only have to specify data for the column up to the last non-null column. See the Examples section for an example of each of these methods of storing the NULL value. NOTE The values of each of the strings specified in the delimiter options must be enclosed by quotation marks. Oracle RMU strips these quotation marks while interpreting the values. If you want to specify a quotation mark (") as a delimiter, specify a string of four quotation marks. Oracle RMU interprets four quotation marks as your request to use one quotation mark as a delimiter. For example, Suffix = """". Oracle RMU reads the quotation marks as follows: o The first quotation mark is stripped from the string. o The second and third quotation marks are interpreted as your request for one quotation mark (") as a delimiter. o The fourth quotation mark is stripped. This results in one quotation mark being used as a delimiter. Furthermore, if you want to specify a quotation mark as part of the delimiter string, you must use two quotation marks for each quotation mark that you want to appear in the string. For example, Suffix = "**""**" causes Oracle RMU to use a delimiter of **"**. A delimiter of blank spaces enclosed in quotes is not valid. o Place_Only=sorted-placement-file Allows you to sort the input file and create an output file sorted in Placement order. The input file can first be sorted into Placement order by using the Place_Only option. The resultant file can then be loaded with the Commit_Every qualifier to gain the required efficiency. Do not use this option with a parallel load operation; parallel load operations perform best when the input file is not sorted. The Place_Only option cannot be used with either the Commit_ Every qualifier or the Exception_File option (data is not being stored in the database). However, the Place_Only option requires the Place qualifier be specified (to sort the data). The placement-sorted output file has the default file extension of .unl. Unless you specify the Null option (with the Format=Delimited_ Text parameter of the Record_Definition qualifier), any null values stored in the rows of the tables being loaded are not preserved. Therefore, use the Null option if you want to preserve null values stored in tables and you are moving data within the database or between databases. See the examples in the Examples help entry under the RMU Unload command for more information. 5 Rms_Record_Def Rms_Record_Def=(File=name[,options]) Rms_Record_Def=(Path=name[.options]) Synonymous with the Record_Definition qualifier. See the description of the Record_Definition qualifier. 5 Restricted_Access Restricted_Access NoRestricted_Access Allows a single process to load data and enables some optimizations available only when restricted access is in use. The default is Norestricted_Access. If you are loading a table from an RMU Unload file which contains LIST OF BYTE VARYING data, the Restricted_Access qualifier reserves the LIST areas for EXCLUSIVE access. This reduces the virtual memory used by long transactions during a load operation and also eliminates I/O to the snapshot files for the LIST storage areas. The Restricted_Access and Parallel qualifiers are mutually exclusive and cannot be specified together on the same RMU Load command line or within a plan file. While RMU Load is running with the Restricted_Access qualifier specified, no other user can attach to the database. 5 Row_Count Row_Count=n Specifies that Oracle Rdb buffer multiple rows between the Oracle Rdb server and the RMU Load process. The default for n is 500 rows; however, this value should be adjusted based on working set size and length of loaded data. Increasing the row count may reduce the CPU cost of the load operation. For remote databases, this may significantly reduce network traffic for large volumes of data because the buffered data can be packaged into larger network packets. The minimum value you can specify for n is 1. The default row size is the value specified for the Commit_Every qualifier or 500, whichever is smaller. 5 Skip Skip=n Noskip Ignores the first n data records in the input file. Use this qualifier in conjunction with the Commit_Every qualifier when restarting an aborted load operation. An aborted load operation displays a message indicating how many records have been committed. Use this value for n. If you specify a negative number, you receive an error message. If you specify a number greater than the number of records in the file, you receive an error message stating that no records have been stored. If you do not specify a value, you receive an error message stating that there is a missing keyword value. Using the Skip qualifier to restart an aborted parallel load operation is rarely useful. Because records are sorted by the controller for each executor involved in the parallel load, there are usually multiple sections of loaded and unloaded records in the input file. Unless you are very familiar with the data you are loading and how it is sorted by the controller, you risk loading some records twice and not loading other records at all, if you use the Skip qualifier when restarting an aborted parallel load operation. The default is the Noskip qualifier. 5 Statistics Statistics=(stat-opts) Specifies that statistics are to be displayed at regular intervals or each time a transaction commits, or both, so that you can evaluate the progress of the load operation. The stat-opts are the options you can specify with this qualifier, namely: Interval=n, On_Commit, or both. If the Statistics qualifier is specified, you must also specify at least one option. When the Statistics=(Interval=n) qualifier is specified, Oracle RMU prints statistics every n seconds. The minimum value for n is 1. When the Statistics=(On_Commit) qualifier is specified, Oracle RMU prints statistics each time a transaction is committed. If you specify both options, Statistics=(Interval=n, On_Commit), statistics are displayed every n seconds and each time a transaction commits. The displayed statistics include: o Elapsed time o CPU time o Buffered I/O o Direct I/O o Page faults o Number of records loaded when the last transaction was committed o Number of records loaded so far in the current transaction o If the Record_Definition=Exception_File option is also specified, the following statistics are displayed also: - Number of records rejected when the last transaction was committed - Number of records rejected so far in the current transaction o If the Parallel qualifier is specified also, the following statistics are displayed also: - Number of extra commits performed by executors Extra commits are caused when the Oracle RMU directs your process or the executors to commit a transaction earlier than usual to avoid a hung load operation. For example, if one executor is holding, but no longer needs a lock that another executor requires, Oracle RMU directs the first executor to commit its current transaction. By directing an executor or executors to commit a transaction earlier than usual, the locks under contention are released and the load operation can proceed. - The total number of executors - The number of executors that are initializing, idle, terminated, sorting, storing, committing, or executing At any time during the load operation, you can press Ctrl/T to display the current statistics. 5 Transaction_Type Transaction_Type=share-mode Specifies the share mode for the load operation. The following share modes are available: Batch_Update Exclusive Protected Shared You must specify a value if you use the Transaction_Type qualifier. If you do not specify the Transaction_Type qualifier, the default share mode is Protected. If you specify a parallel load operation (with the Parallel qualifier), and constraints are defined on the table you are loading, Oracle Corporation recommends that you specify the Shared share mode, or drop the constraints prior to starting a parallel load operation, or specify the Noconstraints qualifier. See the Usage Notes for details. 5 Trigger_Relations Trigger_Relations[=(table-name-list)] NoTrigger_Relations You can use the Trigger_Relations qualifier in three ways: o Trigger_Relations=(table-name-list) Specifies the tables to be reserved for update. Using this qualifier, you can explicitly lock tables that are updated by triggers in store operations. If you list multiple tables, separate the table names with a comma, and enclose the list of table names within parentheses. o Trigger_Relations If you omit the list of table names, the tables updated by triggers are locked automatically as required. This is the default. o NoTrigger_Relations Disables triggers on the target table. This option requires DROP privilege on the table being loaded. You cannot specify a list of table names with this option. If you specify a parallel load operation (with the Parallel qualifier), and triggers are defined on the table you are loading, Oracle Corporation recommends that you specify the Shared share mode or drop the triggers prior to starting a parallel load operation. See the Usage Notes for details. The Trigger_Relations qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 5 Virtual_Fields Virtual_Fields(=[No]Automatic) Novirtual_Fields The Virtual_Fields qualifier is required to reload any AUTOMATIC (or IDENTITY) fields with real data. The Novirtual_Fields qualifier is the default, which is equivalent to the Virtual_Fields=(Noautomatic) qualifier. If you specify the Virtual_Fields qualifier without a keyword, all fields are loaded except COMPUTED BY columns and calculated VIEW columns. Use this qualifier when restructuring a table and when you do not wish the AUTOMATIC INSERT AS or IDENTITY column to recompute new values. Instead, RMU will reload the saved values from a file created by RMU/UNLOAD/VIRTUAL_FIELDS=AUTOMATIC. 4 Usage_Notes o To use the RMU Load command for a database, you must have the RMU$LOAD privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. The appropriate Oracle Rdb privileges for accessing the database tables involved are also required. o To use the RMU Load command with the Audit qualifier, you must have both of the following: - The RMU$SECURITY privilege in the root file ACL for the database whose security audit records are being loaded - The RMU$LOAD privilege in the root file ACL for the database into which these security audit records are being loaded If you do not have both of the privileges described in the preceding list, you must have the OpenVMS SYSPRV or BYPASS privilege. o You can unload a table from a database structured under one version of Oracle Rdb and load it into the same table of a database structured under another version of Rdb. For example, if you unload the EMPLOYEES table from a mf_personnel database created under Oracle Rdb V6.0, you can load the generated .unl file into an Oracle Rdb V7.0 database. Likewise, if you unload the EMPLOYEES table from a mf_personnel database created under Oracle Rdb V7.0, you can load the generated .unl file into an Oracle Rdb V6.1 database. This is true even for specially formatted binary files (created with the RMU Unload command without the Record_Definition qualifier). The earliest version into which you can load a .unl file from another version is Oracle Rdb V6.0. o The following list provides information on parallel load operations: - Specify no more executors (with the Executor_Count option to the Parallel qualifier) than storage areas defined for the table you are loading. - You cannot use a parallel load operation to load list data (segmented string) records or security audit records. If you specify a parallel load operation and attempt to load list data or security audit records, Oracle RMU returns a warning and performs a single-executor load operation. - Oracle Corporation recommends that you specify a shared mode transaction type or specify the Noconstraints qualifier and drop triggers during a parallel load operation; otherwise, constraints and triggers defined on the table you are loading can cause lock conflicts among the parallel load executors. - If you are using parallel load and hashed indexes, do not sort the data prior to loading it. Instead, use the Place qualifier to the RMU Load command to sort the data as it is loaded. (The Place qualifier is useful for hashed indexes, not sorted.) o The following list provides information on loading security audit journals: - Loading security audit journals into a database other than that which is being audited When you load the security audit journals recorded for one database into another database, you specify the database that is being audited as a parameter to the Audit=Database_ File qualifier, and you specify the database into which these security audit records should be loaded with the root-file-spec parameter to the Oracle RMU command. For instance, the following example loads the security audit journal records for the mf_personnel database into the MFP_AUDIT table of the audit_db database. Note that SECURITY_AUDIT is a logical name that points to the actual security audit journal file. $ RMU/LOAD/AUDIT=DATABASE_FILE=MF_PERSONNEL AUDIT_DB - _$ MFP_AUDIT SECURITY_AUDIT When you issue the preceding RMU Load command, the audit_ db database must exist. However, the RMU Load command creates the MFP_AUDIT table in the audit_db database and appropriately defines the columns for the MFP_AUDIT database. In other words, the following SQL statement satisfies the minimum requirements for the audit_db database to be used correctly by the preceding RMU Load command: SQL> CREATE DATABASE FILENAME audit_db.rdb; Note that there is no field in the audit record loaded by Oracle RMU to indicate the source database for the records. Therefore, it is not wise to mix auditing records from different databases in the same table. Instead, auditing information for different databases should be loaded into separate tables. - Security audit journal file name The name of the security audit journal file depends on the version of the operating system software you are running and on the hardware platform, as follows: * SYS$MANAGER:SECURITY.AUDIT$JOURNAL for OpenVMS Alpha V6.1 and later and OpenVMS VAX V6.0 and later * SYS$MANAGER:SECURITY_AUDIT.AUDIT$JOURNAL for OpenVMS Alpha prior to V6.1 and OpenVMS VAX V5.5 and earlier. - Loading security audit journals into the database being audited The Oracle Rdb table into which you load the security audit journal records should be defined with the columns shown in Columns in a Database Table for Storing Security Audit Journal Records under the column marked Oracle Rdb Column Name so that the audit journal records can be loaded successfully into the table. If the table does not exist, the RMU Load Audit command creates it with the columns shown in Columns in a Database Table for Storing Security Audit Journal Records under the column marked Oracle Rdb Column Name. You can give the table any valid name. - Columns in a Database Table for Storing Security Audit Journal Records lists the column names created by the RMU Load command with the Audit qualifier. Table 12 Columns in a Database Table for Storing Security Audit Journal Records Oracle Rdb Column Name SQL Data Type and Length AUDIT$EVENT CHAR 16 AUDIT$SYSTEM_NAME CHAR 15 AUDIT$SYSTEM_ID CHAR 12 AUDIT$TIME_STAMP CHAR 48 AUDIT$PROCESS_ID CHAR 12 AUDIT$USER_NAME CHAR 12 AUDIT$TSN CHAR 25 AUDIT$OBJECT_NAME CHAR 255 AUDIT$OBJECT_TYPE CHAR 12 AUDIT$OPERATION CHAR 32 AUDIT$DESIRED_ CHAR 16 ACCESS AUDIT$SUB_STATUS CHAR 32 AUDIT$FINAL_ CHAR 32 STATUS AUDIT$RDB_PRIV CHAR 16 AUDIT$VMS_PRIV CHAR 16 AUDIT$GRANT_IDENT CHAR 192 AUDIT$NEW_ACE CHAR 192 AUDIT$OLD_ACE CHAR 192 AUDIT$RMU_COMMAND CHAR 512 o Dates stored in ASCII text format can be converted to the VMS DATE data type format by the RMU Load command. See Example 7 in the Examples help entry under this command, which demonstrates this conversion. o To preserve the NULL indicator in a load or unload operation, specify the Null option when you use the Record_Definition qualifier. Using the Record_Definition qualifier without the Null option causes the RMU Load command to replace all NULL values with zeros. This can cause unexpected results with computed-by columns. o When the RMU Load command is issued for a closed database, the command executes without other users being able to attach to the database. o The RMU Load command recognizes character set information. When you load a table, the RMU Load command recognizes that the correct size of a column is based on its character set. For example, the RMU Load command recognizes that a column defined as CHAR (10) CHARACTER SET KANJI occupies 20 octets. o By default, the RMU Load command changes any table or column names that you specify to uppercase. To preserve lowercase characters, use delimited identifiers; that is, enclose the names in quotation marks (""). o If your database uses a character set other than the DEC Multinational character set (MCS) for table and domain names, or if you edit a record definition file to use names from such a character set, the RMU Load command could fail and return the error shown in the following example: $ RMU/UNLOAD/RECORD_DEFINITION=FILE=STRINGS MIA - "TAB_°¡°¢abcd°§ABCD°©°ª" - STRINGS.UNL %RMU-I-DATRECUNL, 4 data records unloaded $ RMU LOAD/RECORD_DEFINITION=FILE=STRINGS MIA - "TAB_°¡°¢abcd°§ABCD°©°ª" - STRINGS.UNL DEFINE FIELD DEC_MCS_CHAR DATATYPE IS TEXT SIZE IS 20. DEFINE FIELD KANJI_CHAR DATATYPE IS TEXT SIZE IS 10 CHARACTERS - CHARACTER SET IS KANJI. DEFINE FIELD HANZI_CHAR DATATYPE IS TEXT SIZE IS 10 CHARACTERS - CHARACTER SET IS HANZI. DEFINE FIELD HANYU_CHAR DATATYPE IS TEXT SIZE IS 10 CHARACTERS - CHARACTER SET IS HANYU. . . . DEFINE RECORD TAB_°¡°¢abcd°§ABCD°©°ª. %RMU-F-RECDEFSYN, Syntax error in record definition file DEFINE RECORD TAB_''°¡°¢ABCD°§ABCD°©°ª. When this problem occurs, edit the record definition file and modify the names so that they can be represented with the MCS character set. o Oracle RMU does not support the multischema naming convention and returns an error if you specify one. For example: $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME) - _$ /RECORD_DEFINITION=(FILE=TEXT_NAMES,EXCEPTION_FILE=FILE.UNL) - _$ corporate_data ADMINISTRATION.PERSONNEL.EMPLOYEES EMP.UNL %RDB-E-BAD_DPB_CONTENT, invalid database parameters in the database parameter block (DPB) %RMU-I-DATRECSTO, 0 data records stored %RMU-I-DATRECREJ, 0 data records rejected. When using a multischema database, you must specify the SQL stored name for the database object. For example, to find the stored name that corresponds to the ADMINISTRATION.PERSONNEL.EMPLOYEES table in the corporate_ data database, issue an SQL SHOW TABLE command. SQL> SHOW TABLE ADMINISTRATION.PERSONNEL.EMPLOYEES Information for table ADMINISTRATION.PERSONNEL.EMPLOYEES Stored name is EMPLOYEES . . . Then, to load the table, issue the following RMU Load command: $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME) - _$ /RECORD_DEFINITION=(FILE=TEXT_NAMES,EXCEPTION_FILE=FILE.UNL) - _$ CORPORATE_DATA EMPLOYEES MY_DATA.UNL %RMU-I-DATRECSTO, 3 data records stored %RMU-I-DATRECREJ, 0 data records rejected. The Fields qualifier can be used with indirect file references. When you use an indirect file reference in the field list, the referenced file is written to SYS$OUTPUT if the DCL SET VERIFY comand has been used. See the Indirect- Command-Files help entry for more information. o The Transaction_Type=Batch_Update qualifier cannot be used with multiple executors (Executor_Count greater than 1). o The RMU Load procedure supports the loading of tables that reference system domains. o If you use a synonym to represent a table or a view, the RMU Load command translates the synonym to the base object and processes the data as though the base table or view had been named. This implies that the unload interchange files (.UNL) or record definition files (.RRD) that contain the table metadata will name the base table or view and not use the synonym name. If the metadata is used against a different database, you may need to use the Match_Name qualifier to override this name during the RMU load process. 4 Examples Example 1 This command loads the data from the RMS file, names.unl, into the newly created RETIREES table of the mf_personnel database. The record structure of RETIREES is in the file names.rrd. The names.unl and names.rrd files were created by a previous RMU Unload command. The unload operation unloaded data from a view derived from a subset of columns in the EMPLOYEES table. $ RMU/LOAD/RECORD_DEFINITION=FILE=NAMES.RRD - _$ MF_PERSONNEL RETIREES NAMES.UNL Example 2 This command restarts an aborted load operation that was loading the newly created RETIREES table of the mf_personnel database from the names.unl file. The columns being loaded are EMPLOYEE_ ID, LAST_NAME, and FIRST_NAME. The original load operation had committed 25 records. Beginning with the 26th record, the restarted load operation commits the transaction at every record until it reaches the original point of failure. $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME, FIRST_NAME) - _$ /COMMIT_EVERY=1/SKIP=25 MF_PERSONNEL RETIREES NAMES.UNL Example 3 This example loads a new table, PENSIONS, into the mf_personnel database by using record definitions located in the data dictionary. This example assumes that you have first defined a temporary view, TEMP_PENSIONS, combining appropriate columns of the EMPLOYEES and SALARY_HISTORY tables. You must also create a permanent table, PENSIONS, into which you will load the data. Unload the TEMP_PENSIONS view by using the RMU Unload command with the Record_Definition=File=name qualifier to create both an .rrd file containing the column definitions and a data.unl file containing the data from the TEMP_PENSIONS view. Load the new record definitions from the pensions.rrd file into the data dictionary by using the @ command at the CDO prompt. Then you can load the data into the PENSIONS table of the mf_personnel database by using the RMU Load command. $ RMU/UNLOAD/RECORD_DEFINITION=FILE=PENSIONS.RRD MF_PERSONNEL - _$ TEMP_PENSIONS DATA.UNL $ DICTIONARY OPERATOR Welcome to CDO V7.0 The CDD/Repository V7.0 User Interface Type HELP for help CDO> @PENSIONS.RRD CDO> EXIT $ RMU/LOAD/RECORD_DEFINITION=PATH=PENSIONS MF_PERSONNEL PENSIONS - _$ DATA.UNL Example 4 The following command loads the audit records for the mf_ personnel database from the security audit journal file into the AUDIT_TABLE table in the mf_personnel database. Note that if the AUDIT_TABLE table does not exist, the RMU Load command with the Audit qualifier creates it with the columns shown in Columns in a Database Table for Storing Security Audit Journal Records. $ RMU/LOAD/AUDIT MF_PERSONNEL AUDIT_TABLE - _$ SYS$MANAGER:SECURITY.AUDIT$JOURNAL %RMU-I-DATRECREAD, 12858 data records read from input file. %RMU-I-DATRECSTO, 27 data records stored. Example 5 The following command loads the audit records for the mf_ personnel database from the security audit journal file into the AUDIT_TABLE table into the audit_db database. Note that the AUDIT_TABLE table is not created when the database is created. In this case, the RMU Load command with the Audit=Database_ File qualifier creates it with the columns shown in Columns in a Database Table for Storing Security Audit Journal Records. $ RMU/LOAD/AUDIT=DATABASE_FILE=MF_PERSONNEL AUDIT_DB AUDIT_TABLE - _$ SYS$MANAGER:SECURITY.AUDIT$JOURNAL Example 6 This example loads a new table, COLLEGES, into the mf_personnel database by using record definitions located in the data dictionary. A commit operation occurs after every record is stored. The Log_Commits qualifier prints a message after each commit operation. $ RMU/LOAD/RECORD_DEFINITION=FILE=COLLEGES.RRD /COMMIT_EVERY=1 - _$ /LOG_COMMIT MF_PERSONNEL COLLEGES RMU.UNL %RMU-I-DATRECSTO, 1 data records stored %RMU-I-DATRECSTO, 2 data records stored %RMU-I-DATRECSTO, 3 data records stored %RMU-I-DATRECSTO, 4 data records stored %RMU-I-DATRECSTO, 4 data records stored $ Example 7 The following example shows how a date stored in the .unl file as 16-character collatable text can be converted to VMS DATE format when loaded into the database by using the RMU Load command. (The form of the .unl date is yyyymmddhhmmsscc, whereas the form of the VMS DATE is dd-mmm-yyyy:hh:mm:ss.cc. In both cases, y is the year, m is the month, d is the day, h is the hour, m is the minute, s is the second, and c is hundredths of a second. However in the .unl format, the month is expressed as an integer, whereas in the VMS DATE format the month is expressed as a 3-character string.) The example assumes that the default SYS$LANGUAGE is ENGLISH. SQL> -- SQL> -- Show the definition of the TEST table, in which the SQL> -- COL1 column is the VMS DATE data type: SQL> -- SQL> SHOW TABLE DATETEST; Columns for table DATETEST: Column Name Data Type Domain ----------- --------- ------ COL1 DATE VMS . . . $ ! $ ! Show the .unl file that will be loaded into the TEST table: $ ! $ TYPE TEST.UNL $ ! 1991060712351212 $ ! $ ! Note that the .rrd file shows a data type of TEXT of 16 $ ! characters. These 16 characters are the number of characters $ ! specified for the date in the test.unl file: $ ! $ TYPE TEST.RRD DEFINE FIELD COL1 DATATYPE IS text size is 16. DEFINE RECORD TEST. COL1 . END TEST RECORD. $ ! $ ! Load the data in test.unl into the DATETEST table: $ ! $ RMU/LOAD/RMS=FILE=TEST.RRD TEST.RDB DATETEST TEST.UNL %RMU-I-DATRECREAD, 1 data records read from input file. %RMU-I-DATRECSTO, 1 data records stored. $ ! $ SQL SQL> ATTACH 'FILENAME TEST'; SQL> SELECT * FROM DATETEST; COL1 7-JUN-1991 12:35:12.12 1 row selected Example 8 The following example shows how a date stored in the .unl file as 22-character collatable text can be converted to TIMESTAMP format when loaded into the database by using the RMU Load command. The correct format for the .unl TIMESTAMP value is yyyy- mm-dd:hh:mm:ss.cc, where y,m,d,h,m,s,and c represent the same elements of the date and time format as described in Example 7. This example also shows the use of an exception file to trap data that cannot be stored. $ ! Create a column in the mf_personnel database with a $ ! TIMESTAMP datatype: $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL.RDB'; SQL> CREATE TABLE NEWTABLE (COL1 TIMESTAMP); SQL> SHOW TABLE (COLUMN) NEWTABLE; Information for table NEWTABLE Columns for table NEWTABLE: Column Name Data Type Domain ----------- --------- ------ COL1 TIMESTAMP(2) SQL> COMMIT; SQL> EXIT $ ! $ ! Create a .unl file with the data you want to load. Note that $ ! the second value is a valid TIMESTAMP specification, the first $ ! value is not. $ ! $ CREATE TEST.UNL 06-14-1991:12:14:14.14 1991-06-14:12:14:14.14 $ ! $ ! Create an .rrd file that defines the TIMESTAMP field $ ! as a TEXT field: $ ! $ CREATE TEST.RRD DEFINE FIELD COL1 DATATYPE IS TEXT SIZE 22. DEFINE RECORD NEWTABLE. COL1. END NEWTABLE RECORD. $ ! $ ! Attempt to load the data in the .unl file. Oracle RMU returns an $ ! error on the first data record because the date was incorrectly $ ! specified. The first record is written to the exception file, $ ! BAD.DAT. $ ! $ RMU/LOAD/RMS=(FILE=TEST.RRD,EXCEPT=BAD.DAT) MF_PERSONNEL.RDB - _$ NEWTABLE TEST.UNL %RMU-I-LOADERR, Error loading row 1. %RDB-E-CONVERT_ERROR, invalid or unsupported data conversion -COSI-F-IVTIME, invalid date or time %RMU-I-DATRECREAD, 2 data records read from input file. %RMU-I-DATRECSTO, 1 data records stored. %RMU-I-DATRECREJ, 1 data records rejected. $ ! $ ! Type BAD.DAT to view the incorrect data record $ ! $ TYPE BAD.DAT 06-14-1991:12:14:14.14 $ ! $ ! Fetch the data record that stored successfully. $ ! $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL.RDB'; SQL> SELECT * FROM NEWTABLE; COL1 1991-06-14:12:14:14.14 1 rows selected Example 9 Using the RMU Load command, you can load a table in a database by placing the fields in a different order in the database than they were in the input file. The jobs.unl file contains the following: 000001000000000190001Rdb Demonstrator DEMO The jobs.rrd file contains the following: DEFINE FIELD J_CODE DATATYPE IS TEXT SIZE IS 4. DEFINE FIELD WAGE_CL DATATYPE IS TEXT SIZE IS 1. DEFINE FIELD J_TITLE DATATYPE IS TEXT SIZE IS 20. DEFINE FIELD MIN_SAL DATATYPE IS TEXT SIZE 10. DEFINE FIELD MAX_SAL DATATYPE IS TEXT SIZE 10. DEFINE RECORD JOBS. MIN_SAL. MAX_SAL. WAGE_CL. J_TITLE. J_CODE. END JOBS RECORD. The JOBS table has the following structure: Columns for table JOBS: Column Name Data Type Domain ----------- --------- ------ JOB_CODE CHAR(4) JOB_CODE_DOM WAGE_CLASS CHAR(1) WAGE_CLASS_DOM JOB_TITLE CHAR(20) JOB_TITLE_DOM MINIMUM_SALARY INTEGER(2) SALARY_DOM MAXIMUM_SALARY INTEGER(2) SALARY_DOM Notice that: o The ordering of the columns is different for the JOBS table in the database and in the input RMS file. o The names in the .rrd file are also different from the names in the database. o The data types of the salary fields are different (Oracle Rdb will do the conversion). To load the RMS file correctly, you must use the following command: $ RMU/LOAD MF_PERSONNEL JOBS JOBS/RMS=FILE=JOBS - _$ /FIELDS=(MINIMUM_SALARY,MAXIMUM_SALARY,WAGE_CLASS,JOB_TITLE, - _$ JOB_CODE) Notice that the Fields qualifier uses the names of the columns in the JOBS table (not the field names in the .rrd file), but in the order of the RMS file. The names in the .rrd file are immaterial. The purpose of the Fields qualifier is to load the first field in the RMS file into the MINIMUM_SALARY column of the JOBS table, load the second field in the RMS file into the MAXIMUM_SALARY column of the JOBS table, and so forth. The results: SQL> SELECT * FROM JOBS WHERE JOB_CODE = 'DEMO'; JOB_CODE WAGE_CLASS JOB_TITLE MINIMUM_SALARY MAXIMUM_SALARY DEMO 1 Rdb Demonstrator $10,000.00 $19,000.00 Example 10 The following example shows the sequence of steps used to sort a file into placement order by using the Place qualifier and the Place_Only option and then to load the file by using the Commit_ Every qualifier: $ RMU/LOAD/PLACE - _$ /RECORD_DEFINITION=(FILE=NAMES.RRD,PLACE_ONLY=PLACED_NAMES) - _$ MF_PERSONNEL EMPLOYEES UNLOADED_NAMES.UNL $ RMU/LOAD/RECORD_DEFINITION=(FILE=NAMES.RRD) - _$ /COMMIT_EVERY=30 MF_PERSONNEL - _$ EMPLOYEES PLACED_NAMES.UNL %RMU-I-DATRECREAD, 100 data records read from input file. %RMU-I-DATRECSTO, 100 data records stored. Example 11 The following example requests that statistics be displayed at a regular interval of every minute. It loads the data from the RMS file, names.unl, into the EMPLOYEES table of the mf_ personnel database. The record structure of EMPLOYEES is in the file names.rrd. The names.rrd file was created by a previous RMU Unload command that unloaded data from a subset of columns in the EMPLOYEES table. $ RMU/LOAD/STATISTICS=(INTERVAL=60) - _$ /RECORD_DEFINITION=(FILE=NAMES) - _$ /FIELDS=(EMPLOYEE_ID, LAST_NAME) - _$ MF_PERSONNEL EMPLOYEES NAMES.UNL Example 12 The following example uses the Exception_File option to the Record_Definition qualifier to tell Oracle RMU the name of the file to hold the exception records. Oracle RMU returns informational messages to alert you to any data records rejected. $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME) - _$ /RECORD_DEFINITION=(FILE=TEXT_NAMES,EXCEPTION_FILE=FILE.UNL) - _$ MF_PERSONNEL EMPLOYEES NAMES.UNL %RMU-I-LOADERR, Error loading row 1. %RDB-E-NO_DUP, index field value already exists; duplicates not allowed for EMPLOYEES_HASH %RMU-I-LOADERR, Error loading row 17. %RDB-E-NO_DUP, index field value already exists; duplicates not allowed for EMPLOYEES_HASH %RMU-I-LOADERR, Error loading row 33. %RDB-E-NO_DUP, index field value already exists; duplicates not allowed for EMPLOYEES_HASH %RMU-I-LOADERR, Error loading row 155. %RDB-E-NO_DUP, index field value already exists; duplicates not allowed for EMPLOYEES_HASH %RMU-I-DATRECREAD, 200 data records read from input file. %RMU-I-DATRECSTO, 196 data records stored. %RMU-I-DATRECREJ, 4 data records rejected. Example 13 The following is an example of the format in which you can provide input data to the RMU Load command when you use the Format=Delimited_Text option with the Record_Definition qualifier. This is followed by the RMU Load command you use to load this data. "99997","ABUSHAKRA","CAROLINE","S","5 CIRCLE STREET","BOX 506", "CHELMSFORD", "MA", "02184", "1960061400000000"# "99996","BRADFORD","LEO","M","4 PLACE STREET","BOX 555", "NASHUA","NH", "03060", "1949051800000000"# $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME, FIRST_NAME, - _$ MIDDLE_INITIAL, ADDRESS_DATA_1, ADDRESS_DATA_2, - _$ CITY, STATE, POSTAL_CODE, BIRTHDAY) - _$ /RECORD_DEFINITION=(FILE= NAMES.RRD, - _$ FORMAT=DELIMITED_TEXT, - _$ TERMINATOR="#" ) - _$ MF_PERSONNEL EMPLOYEES NAMES.UNL %RMU-I-DATRECREAD, 2 data records read from input file. %RMU-I-DATRECSTO, 2 data records stored. Example 14 The following is an example of the format in which you must provide input data to the RMU Load command when you specify the Format=Text option with the Record_Definition qualifier. This is followed by the RMU Load command you use to load this data. 09166Watts Leora F 09190Margolis David M 09187McDonald Lois F $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME, FIRST_NAME, SEX) - _$ /RECORD_DEFINITION=(FILE=TEXT_NAMES.RRD, FORMAT=TEXT) - _$ MF_PERSONNEL EMPLOYEES NAMES.UNL %RMU-I-DATRECREAD, 3 data records read from input file. %RMU-I-DATRECSTO, 3 data records stored. 4 Examples_(Cont.) Example 15 The following example assumes you want to load a data file into the JOBS table that contains more fields than the table definition in the mf_personnel database. The example first attempts to do this by just excluding the extra field from the list associated with the Fields qualifier. However, this causes an error to be returned. The example then uses the FILLER keyword in the .rrd file to tell Oracle RMU not to attempt to load the additional field. The command executes successfully. The table definition for the JOBS table is as follows: Columns for table JOBS: Column Name Data Type Domain ----------- --------- ------ JOB_CODE CHAR(4) JOB_CODE_DOM Primary Key constraint JOBS_PRIMARY_JOB_CODE WAGE_CLASS CHAR(1) WAGE_CLASS_DOM JOB_TITLE CHAR(20) JOB_TITLE_DOM MINIMUM_SALARY INTEGER(2) SALARY_DOM MAXIMUM_SALARY INTEGER(2) SALARY_DOM The .rrd file for the data you want to load appears as follows (note that there is no corresponding field to JOB_STATUS in the mf_personnel database definition for the JOBS table): DEFINE FIELD JOB_CODE DATATYPE IS TEXT SIZE IS 4. DEFINE FIELD WAGE_CLASS DATATYPE IS TEXT SIZE IS 1. DEFINE FIELD JOB_TITLE DATATYPE IS TEXT SIZE IS 20. DEFINE FIELD MINIMUM_SALARY DATATYPE IS TEXT SIZE IS 13. DEFINE FIELD MAXIMUM_SALARY DATATYPE IS TEXT SIZE IS 13. DEFINE FIELD JOB_STATUS DATATYPE IS TEXT SIZE IS 4. DEFINE RECORD JOBS. JOB_CODE . WAGE_CLASS . JOB_TITLE . MINIMUM_SALARY . MAXIMUM_SALARY . JOB_STATUS . END JOBS RECORD. The data file you want to load, jobs.unl, appears as follows: DBAD4Corp Db Administratr55000.00 95000.00 Old You attempt to load the file in the mf_personnel database by listing only the fields in the RMU Load command that have corresponding fields defined in the database: $ RMU/LOAD MF_PERSONNEL/RMS=(FILE=JOBS.RRD, FORMAT=TEXT) - _$ /FIELDS=(JOB_CODE, WAGE_CLASS, JOB_TITLE, MINIMUM_SALARY, - _$ MAXIMUM_SALARY) JOBS JOBS.UNL %RMU-F-FLDMUSMAT, Specified fields must match in number and datatype with the unloaded data %RMU-I-DATRECSTO, 0 data records stored The workaround for the problem of a mismatch between your data and .rrd file, and database definition for a table is to use the FILLER keyword in your .rrd file, as follows: DEFINE FIELD JOB_CODE DATATYPE IS TEXT SIZE IS 4. DEFINE FIELD WAGE_CLASS DATATYPE IS TEXT SIZE IS 1. DEFINE FIELD JOB_TITLE DATATYPE IS TEXT SIZE IS 20. DEFINE FIELD MINIMUM_SALARY DATATYPE IS TEXT SIZE IS 13. DEFINE FIELD MAXIMUM_SALARY DATATYPE IS TEXT SIZE IS 13. DEFINE FIELD JOB_STATUS DATATYPE IS TEXT SIZE IS 4 FILLER. <------ DEFINE RECORD JOBS. JOB_CODE . WAGE_CLASS . JOB_TITLE . MINIMUM_SALARY . MAXIMUM_SALARY . JOB_STATUS . END JOBS RECORD. Now that the .rrd file has been modified, attempt to load the record again: $ RMU/LOAD MF_PERSONNEL/RMS=(FILE=JOBS.RRD, FORMAT=TEXT) - _$ /FIELDS=(JOB_CODE, WAGE_CLASS, JOB_TITLE, MINIMUM_SALARY, - _$ MAXIMUM_SALARY) JOBS JOBS.UNL %RMU-I-DATRECSTO, 1 data records stored. Example 16 The following example demonstrates the use of the Null="*" option of the Record_Definition qualifier to signal to Oracle RMU that any data that appears as an unquoted asterisk in the .unl file should have the corresponding column in the database be flagged as NULL. The example shows the contents of the .unl file, followed by the RMU Load command used to load this .unl file, and then the output from an SQL statement to display the data loaded. "98888","ABUSHAKRA","CAROLINE",*,"5 CIRCLE STREET","BOX 506", "CHELMSFORD", "MA", "02184", "1960061400000000"# "98889","BRADFORD","LEO",*,"4 PLACE STREET","BOX 555", "NASHUA","NH", "03060", "1949051800000000"# $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME, FIRST_NAME, - _$ MIDDLE_INITIAL, ADDRESS_DATA_1, ADDRESS_DATA_2, - _$ CITY, STATE, POSTAL_CODE, BIRTHDAY) - _$ /RECORD_DEFINITION=(FILE= EMPLOYEES.RRD, - _$ FORMAT=DELIMITED_TEXT, - _$ TERMINATOR="#", - -$ NULL="*" ) - _$ MF_PERSONNEL EMPLOYEES EMPLOYEES.UNL %RMU-I-DATRECREAD, 2 data records read from input file. %RMU-I-DATRECSTO, 2 data records stored. SQL> ATTACH 'FILENAME MF_PERSONNEL.RDB'; SQL> SELECT * FROM EMPLOYEES WHERE EMPLOYEE_ID > '98000' cont> AND MIDDLE_INITIAL IS NULL; EMPLOYEE_ID LAST_NAME FIRST_NAME MIDDLE_INITIAL ADDRESS_DATA_1 ADDRESS_DATA_2 CITY STATE POSTAL_CODE SEX BIRTHDAY STATUS_CODE 98888 ABUSHAKRA CAROLINE NULL 5 CIRCLE STREET BOX 506 CHELMSFORD MA 02184 ? 14-Jun-1960 N 98889 BRADFORD LEO NULL 4 PLACE STREET BOX 555 NASHUA NH 03060 ? 18-May-1949 N 2 rows selected Example 17 The following example demonstrates the use of the Null="" option of the Record_Definition qualifier to signal to Oracle RMU that any data that is an empty string in the .unl file (as represented by two commas with no space separating them) should have the corresponding column in the database be flagged as NULL. The example shows the contents of the .unl file, followed by the RMU Load command used to load this .unl file, and then the output from an SQL statement to display the data loaded. "90021","ABUSHAKRA","CAROLINE","A","5 CIRCLE STREET",, "CHELMSFORD", "MA", "02184", "1960061400000000"# "90015","BRADFORD","LEO","B","4 PLACE STREET",, "NASHUA","NH", "03030", "1949051800000000"# $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME, FIRST_NAME, - _$ MIDDLE_INITIAL, ADDRESS_DATA_1, ADDRESS_DATA_2, - _$ CITY, STATE, POSTAL_CODE, BIRTHDAY) - _$ /RECORD_DEFINITION=(FILE= EMPLOYEES.RRD, - _$ FORMAT=DELIMITED_TEXT, - _$ TERMINATOR="#", - _$ NULL="") - _$ MF_PERSONNEL EMPLOYEES EMPLOYEES.UNL %RMU-I-DATRECREAD, 2 data records read from input file. %RMU-I-DATRECSTO, 2 data records stored. $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL.RDB'; SQL> SELECT * FROM EMPLOYEES WHERE ADDRESS_DATA_2 IS NULL; EMPLOYEE_ID LAST_NAME FIRST_NAME MIDDLE_INITIAL ADDRESS_DATA_1 ADDRESS_DATA_2 CITY STATE POSTAL_CODE SEX BIRTHDAY STATUS_CODE 90021 ABUSHAKRA CAROLINE A 5 CIRCLE STREET NULL CHELMSFORD MA 02184 ? 14-Jun-1960 N 90015 BRADFORD LEO B 4 PLACE STREET NULL NASHUA NH 03030 ? 18-May-1949 N 2 rows selected Example 18 The following example is the same as Example 17 except it shows the use of the default value for the Null option of the Record_ Definition qualifier to signal to Oracle RMU that any data that is an empty string in the .unl file (as represented by two commas with no space separating them) should have the corresponding column in the database be flagged as NULL. The example shows the contents of the .unl file, followed by the RMU Load command used to load this .unl file, and then the output from an SQL statement to display the data loaded. "90022","ABUSHAKRA","CAROLINE","A","5 CIRCLE STREET",, "CHELMSFORD", "MA", "02184", "1960061400000000"# "90014","BRADFORD","LEO","B","4 PLACE STREET",, "NASHUA","NH", "03030", "1949051800000000"# $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME, FIRST_NAME, - _$ MIDDLE_INITIAL, ADDRESS_DATA_1, ADDRESS_DATA_2, - _$ CITY, STATE, POSTAL_CODE, BIRTHDAY) - _$ /RECORD_DEFINITION=(FILE= EMPLOYEES.RRD, - _$ FORMAT=DELIMITED_TEXT, - _$ TERMINATOR="#", - _$ NULL) - _$ MF_PERSONNEL EMPLOYEES EMPLOYEES.UNL %RMU-I-DATRECREAD, 2 data records read from input file. %RMU-I-DATRECSTO, 2 data records stored. $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL.RDB'; SQL> SELECT * FROM EMPLOYEES WHERE EMPLOYEE_ID = '90022' OR cont> EMPLOYEE_ID ='90014' AND ADDRESS_DATA_2 IS NULL; EMPLOYEE_ID LAST_NAME FIRST_NAME MIDDLE_INITIAL ADDRESS_DATA_1 ADDRESS_DATA_2 CITY STATE POSTAL_CODE SEX BIRTHDAY STATUS_CODE 90014 BRADFORD LEO B 4 PLACE STREET NULL NASHUA NH 03030 ? 18-May-1949 N 90022 ABUSHAKRA CAROLINE A 5 CIRCLE STREET NULL CHELMSFORD MA 02184 ? 14-Jun-1960 N 2 rows selected Example 19 The following example demonstrates the use of the Null option of the Record_Definition qualifier to signal to Oracle RMU that any data that is an empty string in the .unl file (as represented by two commas with no space separating them) should have the corresponding column in the database be flagged as NULL. In addition, any column for which there is only data for the first column or columns has the remaining columns set to NULL. The example shows the contents of the .unl file, followed by the RMU Load command used to load this .unl file, and then the output from an SQL statement to display the data loaded. "90026","ABUSHAKRA","CAROLINE","A","5 CIRCLE STREET","BOX 783", "CHELMSFORD","MA", "02184", "1960061400000000" "90011","BRADFORD","LEO",,,, "NASHUA","NH","03030","1949051800000000" "90010" "90009",,,,,,,,,"1966061600000000" $ RMU/LOAD/FIELDS=(EMPLOYEE_ID, LAST_NAME, FIRST_NAME, - _$ MIDDLE_INITIAL, ADDRESS_DATA_1, ADDRESS_DATA_2, - _$ CITY, STATE, POSTAL_CODE, BIRTHDAY) - _$ /RECORD_DEFINITION=(FILE= EMPLOYEES.RRD, - _$ FORMAT=DELIMITED_TEXT, - _$ NULL) - _$ MF_PERSONNEL EMPLOYEES EMPLOYEES.UNL %RMU-I-DATRECREAD, 5 data records read from input file. %RMU-I-DATRECSTO, 5 data records stored. $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL.RDB'; SQL> SELECT * FROM EMPLOYEES WHERE EMPLOYEE_ID ='90026' OR cont> EMPLOYEE_ID BETWEEN '90009' AND '90011'; EMPLOYEE_ID LAST_NAME FIRST_NAME MIDDLE_INITIAL ADDRESS_DATA_1 ADDRESS_DATA_2 CITY STATE POSTAL_CODE SEX BIRTHDAY STATUS_CODE 90009 NULL NULL NULL NULL NULL NULL NULL NULL ? 16-Jun-1966 N 90010 NULL NULL NULL NULL NULL NULL NULL NULL ? NULL N 90011 BRADFORD LEO NULL NULL NULL NASHUA NH 03030 ? 18-May-1949 N 90026 ABUSHAKRA CAROLINE A 5 CIRCLE STREET BOX 783 CHELMSFORD MA NULL ? 14-Jun-1960 N 4 rows selected Example 20 The following example demonstrates a parallel load operation. In this example, three executors are specified because there are three storage areas in the JOB_HISTORY table of the mf_personnel database. The Defer_Index_Updates qualifier is used because there are no constraints or triggers defined on the JOB_HISTORY table, and it is known that no other database activity will occur when this command is executed. In addition, a plan file is generated to capture the specification of this load operation. See the next example for a description of the plan file. Note that the pid provided in the output from this command is the process ID. $ RMU/LOAD/PARALLEL=(EXEC=3)/DEFER_INDEX_UPDATES mf_personnel.rdb - _$ /RECORD_DEFINITION=(FILE=JOB_HIST,FORMAT=DELIMITED_TEXT, - _$ EXCEPTION_FILE=DISK1:[ERRORS]JOB_HIST.EXC) - _$ /STATISTICS=(INTERVAL=30)/LIST_PLAN=JOB_HISTORY.PLAN - _$ JOB_HISTORY JOB_HIST.UNL %RMU-I-EXECUTORMAP, Executor EXECUTOR_1 (pid: 2941941B) will load storage area EMPIDS_LOW. %RMU-I-EXECUTORMAP, Executor EXECUTOR_2 (pid: 2941F01D) will load storage area EMPIDS_MID. %RMU-I-EXECUTORMAP, Executor EXECUTOR_3 (pid: 2941C81F) will load storage area EMPIDS_OVER. -------------------------------------------------------------------------- ELAPSED: 0 00:00:30.05 CPU: 0:00:01.64 BUFIO: 59 DIRIO: 219 FAULTS: 2670 1640 data records read from input file. 1330 records loaded before last commit. 220 records loaded in current transaction. 0 records rejected before last commit. 0 records rejected in current transaction. 26 early commits by executors. 3 executors: 0 Initializing; 0 Idle; 0 Terminated 0 Sorting; 2 Storing; 1 Committing; 0 Executing -------------------------------------------------------------------------- . . . -------------------------------------------------------------------------- ELAPSED: 0 00:02:30.12 CPU: 0:00:02.94 BUFIO: 103 DIRIO: 227 FAULTS: 267 1 8070 data records read from input file. 7800 records loaded before last commit. 210 records loaded in current transaction. 0 records rejected before last commit. 0 records rejected in current transaction. 139 early commits by executors. 3 executors: 0 Initializing; 0 Idle; 0 Terminated 0 Sorting; 1 Storing; 2 Committing; 0 Executing --------------------------------------------------------------------------- %RMU-I-EXECSTAT0, Statistics for EXECUTOR_1: %RMU-I-EXECSTAT1, Elapsed time: 00:02:45.84 CPU time: 12.95 %RMU-I-EXECSTAT2, Storing time: 00:00:45.99 Rows stored: 2440 %RMU-I-EXECSTAT3, Commit time: 00:01:33.17 Direct I/O: 6623 %RMU-I-EXECSTAT4, Idle time: 00:00:22.34 Early commits: 47 %RMU-I-EXECSTAT0, Statistics for EXECUTOR_2: %RMU-I-EXECSTAT1, Elapsed time: 00:02:48.42 CPU time: 18.10 %RMU-I-EXECSTAT2, Storing time: 00:01:24.98 Rows stored: 4319 %RMU-I-EXECSTAT3, Commit time: 00:01:18.13 Direct I/O: 9621 %RMU-I-EXECSTAT4, Idle time: 00:00:01.03 Early commits: 29 %RMU-I-EXECSTAT0, Statistics for EXECUTOR_3: %RMU-I-EXECSTAT1, Elapsed time: 00:02:46.50 CPU time: 9.78 %RMU-I-EXECSTAT2, Storing time: 00:00:11.12 Rows stored: 2293 %RMU-I-EXECSTAT3, Commit time: 00:02:26.67 Direct I/O: 3101 %RMU-I-EXECSTAT4, Idle time: 00:00:04.14 Early commits: 77 %RMU-I-EXECSTAT5, Main process idle time: 00:02:41.06 %RMU-I-DATRECREAD, 9052 data records read from input file. %RMU-I-DATRECSTO, 9052 data records stored. %RMU-I-DATRECREJ, 0 data records rejected. Example 21 The following command is the same as in the previous example, except the Noexecute qualifier is specified. Because this qualifier is specified, the load operation is not performed. However, the load plan file is created and verified. $ RMU/LOAD/PARALLEL=(EXEC=3)/DEFER_INDEX_UPDATES/NOEXECUTE - _$ mf_personnel.rdb - _$ /RECORD_DEFINITION=(FILE=JOB_HIST,FORMAT=DELIMITED_TEXT, - _$ EXCEPTION_FILE=DISK1:[ERRORS]JOB_HIST.EXC) - _$ /STATISTICS=(INTERVAL=30)/LIST_PLAN=JOB_HISTORY.PLAN - _$ JOB_HISTORY JOB_HIST.UNL Example 22 The following display shows the contents of the plan file, JOB_HISTORY.PLAN, created in the preceding example. The following callouts are keyed to this display: 1 The Plan Parameters include all the parameters specified on the RMU Load command line and all possible command qualifiers. 2 Command qualifiers that are not specified on the command line are sometimes represented as comments in the plan file. This allows you to edit and adjust the plan file for future use. 3 Command qualifiers that are not specified on the command line and for which there are defaults are sometimes represented with their default value in the plan file. 4 Command qualifiers that are explicitly specified on the command line are represented in the plan file as specified. 5 Executor Parameters are listed for each executor involved in the load operation. Like the command qualifiers, both the values you specify on the command line and those that are allowed but were not specified are included in this list of parameters. 6 Note that the exception file extension is appended with the executor number. When you specify such files on the command line, Oracle RMU generates a separate file for each executor. If desired, you could edit this plan file to place each exception file on a different disk or directory. ! Plan created on 20-JUL-1995 by RMU/LOAD. Plan Name = LOAD_PLAN Plan Type = LOAD Plan Parameters:1 Database Root File = MF_PERSONNEL.RDB; Table Name = JOB_HISTORY Input File = JOB_HIST.UNL ! Fields = 2 Transaction_Type = PROTECTED ! Buffers = Row_Count = 50 3 ! Skip = NoLog_Commits NoCorresponding Defer_Index_Updates Constraints Parallel NoPlace Statistics = INTERVAL = 30 4 NoTrigger_Relations Record_Definition_File = JOB_HIST Format = Delimited_Text Prefix = """" Suffix = """" NoNull Separator = "," End Of Line Terminator End Plan Parameters Executor Parameters: 5 Executor Name = EXECUTOR_1 ! Place_Only = Exception_File = DISK1:[DATABASE]JOB_HIST.EXC_1; 6 ! RUJ Directory = Communication Buffers = 4 End Executor Parameters Executor Parameters: Executor Name = EXECUTOR_2 ! Place_Only = Exception_File = DISK1:[DATABASE]JOB_HIST.EXC_2; ! RUJ Directory = Communication Buffers = 4 End Executor Parameters Executor Parameters: Executor Name = EXECUTOR_3 ! Place_Only = Exception_File = DISK1:[DATABASE]JOB_HIST.EXC_3; ! RUJ Directory = Communication Buffers = 4 End Executor Parameters Example 23 The following example demonstrates the structure of the record definition file (.rrd) for an RMU Load command for several different data types. The first part of the example displays the table definition, the second part shows the RMU Unload command you could use to get an appropriate .rrd file for these data types, and the last part shows the .rrd file definitions for these data types: SQL> attach 'filename data_types.rdb'; SQL> show table many_types; Information for table MANY_TYPES Columns for table MANY_TYPES: Column Name Data Type Domain ----------- --------- ------ F_ID TINYINT F_CHAR_3 CHAR(3) F_TINYINT TINYINT F_SMALLINT SMALLINT F_INTEGER INTEGER F_BIGINT BIGINT F_NTINYINT TINYINT(1) F_NSMALLINT SMALLINT(2) F_NINTEGER INTEGER(7) F_NBIGINT BIGINT(5) F_REAL REAL F_DOUBLE_PREC DOUBLE PRECISION F_DATE_VMS DATE VMS F_DATE_ANSI DATE ANSI F_VARCHAR VARCHAR(20) F_FLOAT REAL F_DATE DATE VMS F_TIME TIME F_TIMESTAMP TIMESTAMP(2) F_INTERVAL INTERVAL DAY (2) $ RMU/UNLOAD DATA_TYPES.RDB/RECORD_DEF=(FILE=MANY_TYPES.RRD) - _$ MANY_TYPES MANY_TYPES.UNL $ TYPE MANY_TYPES.RRD DEFINE FIELD F_ID DATATYPE IS SIGNED BYTE. DEFINE FIELD F_CHAR_3 DATATYPE IS TEXT SIZE IS 3. DEFINE FIELD F_TINYINT DATATYPE IS SIGNED BYTE. DEFINE FIELD F_SMALLINT DATATYPE IS SIGNED WORD. DEFINE FIELD F_INTEGER DATATYPE IS SIGNED LONGWORD. DEFINE FIELD F_BIGINT DATATYPE IS SIGNED QUADWORD. DEFINE FIELD F_NTINYINT DATATYPE IS SIGNED BYTE SCALE -1. DEFINE FIELD F_NSMALLINT DATATYPE IS SIGNED WORD SCALE -2. DEFINE FIELD F_NINTEGER DATATYPE IS SIGNED LONGWORD SCALE -7. DEFINE FIELD F_NBIGINT DATATYPE IS SIGNED QUADWORD SCALE -5. DEFINE FIELD F_REAL DATATYPE IS F_FLOATING. DEFINE FIELD F_DOUBLE_PREC DATATYPE IS G_FLOATING. DEFINE FIELD F_DATE_VMS DATATYPE IS DATE. DEFINE FIELD F_DATE_ANSI DATATYPE IS DATE ANSI. DEFINE FIELD F_VARCHAR DATATYPE IS TEXT SIZE IS 20. DEFINE FIELD F_FLOAT DATATYPE IS F_FLOATING. DEFINE FIELD F_DATE DATATYPE IS DATE. DEFINE FIELD F_TIME DATATYPE IS TIME. DEFINE FIELD F_TIMESTAMP DATATYPE IS TIMESTAMP SCALE -2. DEFINE FIELD F_INTERVAL DATATYPE IS INTERVAL DAY SIZE IS 2 DIGITS. DEFINE RECORD MANY_TYPES. F_ID . F_CHAR_1 . . . . END MANY_TYPES RECORD. Example 24 The following example shows part of a script for loading a copy of the PERSONNEL database using the output from SQL EXPORT. $! Export the database definition and the data $ sql$ export database filename personnel into pers.rbr; $ $! Create an empty database (use RMU Load to add data) $ sql$ import database from pers.rbr filename copy_pers no data; $ $! Now use load to add the same table $ rmu/load copy_pers /match_name=employees employees pers.rbr %RMU-I-DATRECREAD, 100 data records read from input file. %RMU-I-DATRECSTO, 100 data records stored. $ $ rmu/load copy_pers /match_name job_history pers.rbr %RMU-I-DATRECREAD, 274 data records read from input file. %RMU-I-DATRECSTO, 274 data records stored. $ $ rmu/load copy_pers /match_name salary_history pers.rbr %RMU-I-DATRECREAD, 729 data records read from input file. %RMU-I-DATRECSTO, 729 data records stored. $ . . . $ rmu/load copy_pers /match_name work_status pers.rbr %RMU-I-DATRECREAD, 3 data records read from input file. %RMU-I-DATRECSTO, 3 data records stored. Example 25 The following example shows that, by default, truncation errors during a Load are reported. $ rmu/load abc f2 f1 %RMU-I-LOADERR, Error loading row 1. %RDB-E-TRUN_STORE, string truncated during assignment to a column %RMU-I-DATRECREAD, 1 data records read from input file. %RMU-I-DATRECSTO, 0 data records stored. %RMU-F-FTL_LOAD, Fatal error for LOAD operation at 13-FEB-2008 15:39:44.40 $ Example 26 The following example shows the use of the /VIRTUAL_FIELDS qualifier. The values of the INTEGER field A and the AUTOMATIC field B are first unloaded into the AA.UNL file from the RMU_ LOAD_AUTOMATIC_4_DB database table AA using the /VIRTUAL_ FIELDS qualifier. Then the values of the INTEGER field A and the AUTOMATIC field B in the AA.UNL file are loaded into the AA table in the RMU_LOAD_AUTOMATIC_4_DB2 database. $ SQL create database filename RMU_LOAD_AUTOMATIC_4_DB; -- create a sequence and a table create sequence S increment by -1; create table AA (a integer ,b automatic as s.nextval); -- load 10 rows begin declare :i integer; for :i in 1 to 10 do insert into AA (a) values (:i); end for; end; commit; disconnect all; $ exit $ rmu/unload- /virtual=(automatic)- /record=(file=rr,format=delim)- RMU_LOAD_AUTOMATIC_4_DB aa aa.unl %RMU-I-DATRECUNL, 10 data records unloaded. $ $ $! Load using /VIRTUAL $ rmu/load- /record=(file=rr,format=delim)- /virtual- RMU_LOAD_AUTOMATIC_4_DB2 aa aa.unl %RMU-I-DATRECREAD, 10 data records read from input file. %RMU-I-DATRECSTO, 10 data records stored. $ 3 Plan Executes a load plan file previously created with the RMU Load command (or created manually by the user). 4 Description A load plan file is created when you execute an RMU Load command with the List_Plan qualifier. See Load Database for details on creating a plan file, the format of a plan file, and understanding the informational messages returned by a Parallel Load operation. 4 Format (B)0RMU/Load/Plan plan-file Command Qualifiers x Defaults x /[No]Execute x Execute /List_Plan=output-file x None x 4 Parameters 5 plan-file-spec The file specification for the load plan file. The default file extension is .plan. 4 Command_Qualifiers 5 Execute Execute Noexecute The Execute qualifier specifies that the plan file is to be executed. The Noexecute qualifier specifies that the plan file should not be executed, but rather that a validity check be performed on the contents of the plan file. The validity check determines such things as whether the specified table is in the specified database, the .rrd file (if specified) matches the table, and so on. The validity check does not determine such things as whether your process and global page quotas are sufficient. By default, data is loaded when the RMU Load Plan command is issued. 5 List_Plan List_Plan=output-file Specifies that Oracle RMU should generate a new plan file and write it to the specified output file. This new plan file is identical to the plan file you specified on the command line (the "original" plan file) with the following exceptions: o Any comments that appear in the original plan file will not appear in the new plan file. o If the number of executors specified in the original plan file exceeds the number of storage areas that the table being loaded contains, the new plan file will reduce the number of executors to match the number of storage areas. 4 Usage_Notes o To use the RMU Load Plan command for a database, you must have the RMU$LOAD privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. Privileges for accessing the database tables involved are also required. o When the load plan is executed, executors are created as detached processes if you have the OpenVMS DETACH privilege. If you do not have the OpenVMS DETACH privilege, executors are created as subprocesses of your process. 4 Examples Example 1 The following example demonstrates the following: 1. The first Oracle RMU command creates a parallel load plan file. The RMU Load command is not executed because the point of issuing the command is to create the plan file, not to load data. Notice that the created load plan has only three executors, even though four were specified on the command line. This is because EMPLOYEES has only three storage areas. 2. The load plan file generated by the first Oracle RMU command is displayed. 3. The load plan file is edited to change some parameters and to rename the executors with names that describe the storage area each executor is responsible for loading. 4. The edited version of the load plan file is executed. $ RMU/LOAD/PARALLEL=(EXECUTOR_COUNT=4, BUFFER_COUNT=4)/NOEXECUTE - _$ /RECORD_DEFINITION=(FILE=EMPLOYEES.RRD, FORMAT=DELIMITED) - _$ /LIST_PLAN=EMPLOYEES.PLAN MF_PERSONNEL.RDB EMPLOYEES EMPLOYEES.UNL %RMU-W-TOOMANYEXECS, 4 executors were requested, but only 3 executors will be used. $ ! $ TYPE EMPLOYEES.PLAN ! Plan created on 20-JUL-1995 by RMU/LOAD. Plan Name = LOAD_PLAN Plan Type = LOAD Plan Parameters: Database Root File = MF_PERSONNEL.RDB Table Name = EMPLOYEES Input File = EMPLOYEES.UNL ! Fields = Transaction_Type = PROTECTED ! Buffers = Row_Count = 50 ! Skip = NoLog_Commits NoCorresponding NoDefer_Index_Updates Constraints Parallel NoPlace ! Statistics = NoTrigger_Relations Record_Definition_File = EMPLOYEES.RRD Format = Delimited_Text Prefix = """" Suffix = """" NoNull Separator = "," End Of Line Terminator End Plan Parameters Executor Parameters: Executor Name = EXECUTOR_1 ! Place_Only = ! Exception_File = ! RUJ Directory = Communication Buffers = 4 End Executor Parameters Executor Parameters: Executor Name = EXECUTOR_2 ! Place_Only = ! Exception_File = ! RUJ Directory = Communication Buffers = 4 End Executor Parameters Executor Parameters: Executor Name = EXECUTOR_3 ! Place_Only = ! Exception_File = ! RUJ Directory = Communication Buffers = 4 End Executor Parameters The following is an edited version of the plan file presented in the previous example. The file has been edited as follows: o Comments have been added to indicate that the file has been edited. o The Row_Count value has been changed from 50 to 60. o Each executor name has been changed to reflect the storage area the executor is responsible for loading. This makes it easier to determine the storage area from which a record was rejected if an error occurs during loading. In addition, it makes it easier to determine, when records are rejected, which executor was attempting to load it and which Rdb error corresponds to a particular executor. o The directory and file name for each exception file has been changed and the comment character preceding "Exception_File" has been removed. o Directories for the .ruj files have been added and the comment character preceding "RUJ Directory" has been removed. ! Plan created on 20-JUL-1995 by RMU/LOAD. ! Edited on 21-JUL-1995 by John Stuart Plan Name = LOAD_PLAN Plan Type = LOAD Plan Parameters: Database Root File = MF_PERSONNEL.RDB Table Name = EMPLOYEES Input File = EMPLOYEES.UNL ! Fields = Transaction_Type = PROTECTED ! Buffers = Row_Count = 20 ! Skip = NoLog_Commits NoCorresponding NoDefer_Index_Updates Constraints Parallel NoPlace ! Statistics = NoTrigger_Relations Record_Definition_File = EMPLOYEES.RRD Format = Delimited_Text Prefix = """" Suffix = """" NoNull Separator = "," End Of Line Terminator End Plan Parameters Executor Parameters: Executor Name = EMPIDS_LOW_EXEC ! Place_Only = Exception_File = DISK1:[EXCEPTIONS]EMPIDS_LOW.EXC RUJ Directory = DISK1:[RUJ]EMPIDS_LOW.RUJ Communication Buffers = 4 End Executor Parameters Executor Parameters: Executor Name = EMPIDS_MID_EXEC ! Place_Only = Exception_File = DISK2:[EXCEPTIONS]EMPIDS_MID.EXC RUJ Directory = DISK2:[RUJ]EMPIDS_MID.RUJ Communication Buffers = 4 End Executor Parameters Executor Parameters: Executor Name = EMPIDS_OVER_EXEC ! Place_Only = Exception_File = DISK3:[EXCEPTIONS]EMPIDS_LOW.EXC RUJ Directory = DISK3:[RUJ]EMPIDS_LOW.RUJ Communication Buffers = 4 End Executor Parameters $ ! $ ! Execute the plan file. $ ! Each executor is assigned the storage area or areas and $ ! the pid (process ID) for each executor is displayed. $ ! Notice that Oracle RMU notifies you if an error occurs when $ ! an executor attempts to load a row, and then lists the Rdb error $ ! message. Sometimes you receive two or more Oracle RMU $ ! messages in a row and then the associated Oracle Rdb message. You $ ! can match the Oracle RMU message to the Oracle Rdb message by $ ! matching the executor name prefixes to the messages. $ ! $ RMU/LOAD/PLAN EMPLOYEES.PLAN %RMU-I-EXECUTORMAP, Executor EMPIDS_LOW_EXEC (pid: 3140A4CC) will load storage area EMPIDS_LOW. %RMU-I-EXECUTORMAP, Executor EMPIDS_MID_EXEC (pid: 314086CD) will load storage area EMPIDS_MID. %RMU-I-EXECUTORMAP, Executor EMPIDS_OVER_EXEC (pid: 314098CE) will load storage area EMPIDS_OVER. EMPIDS_MID_EXEC: %RMU-I-LOADERR, Error loading row 4. EMPIDS_LOW_EXEC: %RMU-I-LOADERR, Error loading row 1. EMPIDS_MID_EXEC: %RDB-E-NO_DUP, index field value already exists; duplicates not allowed for EMPLOYEES_HASH EMPIDS_LOW_EXEC: %RDB-E-NO_DUP, index field value already exists; duplicates not allowed for EMPLOYEES_HASH %RMU-I-EXECSTAT0, Statistics for EMPIDS_LOW_EXEC: %RMU-I-EXECSTAT1, Elapsed time: 00:00:51.69 CPU time: 4.51 %RMU-I-EXECSTAT2, Storing time: 00:00:32.33 Rows stored: 161 %RMU-I-EXECSTAT3, Commit time: 00:00:00.66 Direct I/O: 932 %RMU-I-EXECSTAT4, Idle time: 00:01:44.99 Early commits: 1 %RMU-I-EXECSTAT0, Statistics for EMPIDS_MID_EXEC: %RMU-I-EXECSTAT1, Elapsed time: 00:01:06.47 CPU time: 4.32 %RMU-I-EXECSTAT2, Storing time: 00:00:38.80 Rows stored: 142 %RMU-I-EXECSTAT3, Commit time: 00:00:01.04 Direct I/O: 953 %RMU-I-EXECSTAT4, Idle time: 00:00:18.18 Early commits: 2 %RMU-I-EXECSTAT0, Statistics for EMPIDS_OVER_EXEC: %RMU-I-EXECSTAT1, Elapsed time: 00:01:04.98 CPU time: 3.22 %RMU-I-EXECSTAT2, Storing time: 00:00:30.89 Rows stored: 100 %RMU-I-EXECSTAT3, Commit time: 00:00:00.90 Direct I/O: 510 %RMU-I-EXECSTAT4, Idle time: 00:00:26.65 Early commits: 1 %RMU-I-EXECSTAT5, Main process idle time: 00:00:58.11 %RMU-I-DATRECREAD, 495 data records read from input file. %RMU-I-DATRECSTO, 403 data records stored. %RMU-I-DATRECREJ, 92 data records rejected. 2 Monitor The Oracle RMU Monitor controls the Oracle Rdb Monitor Process. An Oracle Rdb Monitor Process must be running on each system on which you use Oracle Rdb, including each node in a VAXcluster or VMScluster. An RMU Monitor command controls only the monitor process running on the system from which the command is issued. The Oracle Rdb Monitor Process controls all database access and initiates the automatic recovery procedure when necessary. 3 Reopen_Log Closes the current Oracle Rdb monitor log file, compresses it, and opens another one without stopping the monitor. 4 Description The RMU Monitor Reopen_Log command closes the current Oracle Rdb monitor log file, compresses it, and opens another log file without stopping the monitor. The new log has the same name as, but a new version number of, the monitor log file you opened with the RMU Monitor Start command. Use the RMU Show Users command to determine the current name and location of the monitor log file before issuing the RMU Monitor Reopen_Log command. You should use the RMU Monitor Reopen_Log command if the monitor log file gets too large. For example, if you are running out of space on your disk or if database performance slows, you might want to open another log file. If the disk that contains the Oracle Rdb monitor log file becomes full, you must acquire space on the disk. Once there is sufficient space on this disk, use the RMU Monitor Reopen_Log command and consider backing up (using the DCL COPY command or the OpenVMS Backup utility) the old monitor log file. When the disk that contains the monitor log becomes full, Oracle Rdb stops writing to the log file, but the Oracle Rdb system does not stop operating. A message is sent to the cluster system operator when this occurs. 4 Format (B)0 RMU/Monitor Reopen_Log 4 Usage_Notes o To use the RMU Monitor Reopen_Log command, either you must have the OpenVMS SETPRV privilege or the OpenVMS WORLD, CMKRNL, DETACH, PSWAPM, ALTPRI, SYSGBL, SYSNAM, SYSPRV, and BYPASS privileges. 4 Examples Example 1 The following example closes the existing monitor log file, compresses it, and creates a new one without stopping the Oracle Rdb monitor: $ RMU/MONITOR REOPEN_LOG See the Oracle Rdb Guide to Database Maintenance for more examples that show the RMU Monitor commands. 3 Start Activates the Oracle Rdb monitor process. 4 Description The RMU Monitor Start command activates the Oracle Rdb monitor process (RDMS_MONITORnn, where nn represents the version Oracle Rdb), sets the priority of this process, and specifies a device, directory and file name in which to create the monitor log file. If the monitor process is active already, you receive the following error message: %RMU-F-MONMBXOPN, monitor is already running An Oracle Rdb monitor process must be running on a node for users logged in to that node to use any Oracle Rdb database. In a VMScluster environment, a monitor process must be running on each node in the cluster from which databases are accessed. The Oracle Rdb monitor process controls all database access and initiates the automatic database recovery procedure following a system failure or other abnormal termination of a database user process. See the Oracle Rdb Installation and Configuration Guide for information on support for multiple versions of Oracle Rdb. 4 Format (B)0 RMU/Monitor Start Command Qualifiers x Defaults x /Output = file-name x /Output=SYS$SYSTEM:RDMMON.LOG /Priority = integer x /Priority = 15 /[No]Swap x /Noswap 4 Command_Qualifiers 5 Output Output=file-name Specifies the device, directory, and file name that receives the monitor log. You can use this qualifier to redirect the placement of your monitor log file. The default device and directory is the SYS$SYSTEM directory. The default log file name is RDMMON.LOG. The RMU Monitor Start command causes a new version of the log file to be created for each database session. 5 Priority Priority=integer Specifies the base priority of the monitor process. This priority should always be higher than the highest database user process priority. By default, the monitor runs at the highest interactive priority possible, 15. You should not normally have to lower the monitor process priority. If you change this to a lower priority, an attach operation can cause a deadlock. Deadlock occurs when multiple processes with higher priority than the monitor attempt to attach at the same time. In this case, the monitor must contend for CPU time with multiple higher-priority processes and is perpetually locked out. As a result, no one can use the database. 5 Swap Swap Noswap Enables or disables swapping of the monitor process. The default is Noswap. The Swap qualifier is not recommended for time- critical applications, because no one can use the database while the monitor process is being swapped. 4 Usage_Notes o To use the RMU Monitor Start command, you must have either the OpenVMS SETPRV privilege or the OpenVMS WORLD, CMKRNL, DETACH, PSWAPM, ALTPRI, PRMMBX, SYSGBL, SYSNAM, SYSPRV, and BYPASS privileges. o If the monitor has not been started on the system previously, use the RMONSTART.COM command file (which, by default, is located in the SYS$STARTUP directory) instead of the RMU Monitor Start command. o Start the monitor from the SYSTEM account, which has the SETPRV privilege. The process starting the monitor attempts to give RDMS_MONITOR all privileges. In particular, the privileges required are ALTPRI, CMKRNL, DETACH, PSWAPM, PRMMBX, SETPRV, SYSGBL, SYSNAM, and WORLD. o The monitor process inherits some quotas, such as MAXDETACH, and the user name of the user who starts it. This can result in severe restrictions on user access. For example, if the user who starts the monitor has a MAXDETACH quota of two, then the monitor can only start two recovery processes at one time. However, the system defines most of the quotas needed by the monitor. o If the LNM$PERMANENT_MAILBOX table is not defined in the LNM$SYSTEM_TABLE logical name table, either of the following might occur: - The RMU Start Monitor command hangs - You receive the error, "monitor is not running", when you know it is. By default, the LNM$PERMANENT_MAILBOX table is defined in the LNM$SYSTEM_TABLE logical name table. However, sometimes a user or third-party application redefines the LNM$PERMANENT_MAILBOX table in another logical name table (such as the LNM$GROUP table). To recover from this situation, follow these steps: 1. Define the LNM$PERMANENT_MAILBOX table in the LNM$SYSTEM_TABLE: $ DEFINE/TABLE=LNM$PROCESS_DIRECTORY LNM$PERMANENT_MAILBOX - _$ LNM$SYSTEM 2. Start the database monitor: RMU/MONITOR START 3. Start the application Or, change the application that redefines the LNM$PERMANENT_ MAILBOX table so that LNM$PERMANENT_MAILBOX is defined as a search list that includes the LNM$SYSTEM_TABLE table, as shown in the following example: $ DEFINE/TABLE=LNM$PROCESS_DIRECTORY LNM$PERMANENT_MAILBOX - _$ LNM$GROUP, LNM$SYSTEM o Use the RMU Show System command to determine the location of the monitor log file if it is not in the default location. The monitor log file may not be in the default location if someone has issued the RMU Monitor Start command and specified a location different from the default with the Output qualifier. o The monitor process should only be started by a user whose account has adequate quotas. Ideally, the monitor process should be started from the SYSTEM account. o To view the contents of monitor log file online (even when disk-based logging is disabled because of disk space problems), use the Performance Monitor and select the Monitor Log screen from the Per-Process menu. See the Oracle Rdb7 Guide to Database Performance and Tuning or the Performance Monitor Help for information about using the Performance Monitor. 4 Examples Example 1 The following command activates the Oracle Rdb monitor process: $ RMU/MONITOR START See the Oracle Rdb Guide to Database Maintenance for more examples that show the RMU Monitor commands. 3 Stop Stops the Oracle Rdb monitor process. 4 Description The RMU Monitor Stop command stops the Oracle Rdb monitor process (RDMS_MONITORnn, where nn represents the version Oracle Rdb) normally, either with a shutdown and rollback of the databases or an immediate abort. You can use the RMU Monitor Stop command to shut down all database activity on your node, optionally aborting user processes by forcing an image exit or deleting their processes. The RMU Monitor Stop command closes the monitor log file also. An Oracle Rdb monitor process must be running on a node for users logged in to that node to use any Oracle Rdb database. In a VMScluster environment, a monitor process must be running on each node in the cluster from which databases is accessed. The Oracle Rdb monitor process controls all database access and initiates the automatic database recovery procedure following a system failure or other abnormal termination of a database user process. The monitor log file automatically tracks all access to the database. 4 Format (B)0 RMU/Monitor Stop Command Qualifiers x Defaults x /[No]Abort[={Forcex | Delprc}] x /NOABORT /[No]Wait x /NOWAIT 4 Command_Qualifiers 5 Abort Abort=Forcex Abort=Delprc Noabort The Abort=Forcex qualifier stops the monitor immediately without allowing current Oracle Rdb users to complete active transactions or detach from their databases. However, the user processes are not deleted. Active transactions are rolled back. If a process using a database is waiting for a subprocess to complete, the transaction is not rolled back until the subprocess completes. Using the Abort qualifier with no option is equivalent to specifying the Abort=Forcex qualifier. The Abort=Delprc qualifier stops the monitor immediately without allowing current Oracle Rdb users to complete active transactions or detach from their databases. Each user process that was attached to an Oracle Rdb database is deleted immediately. The Noabort qualifier allows current user processes to continue and complete before stopping. New users on the node are not allowed to attach to any database, but existing database users can complete their sessions normally. Once existing database user processes terminate, the database monitor shuts down. The Noabort qualifier is the default. 5 Wait Wait Nowait Specifies whether the Oracle RMU operation completes when the monitor acknowledges the stop request (Nowait), or whether RMU waits until the monitor finishes shutting down (Wait). The default is Nowait. 4 Usage_Notes o To use the RMU Monitor Stop command, you must have either the OpenVMS SETPRV privilege or the OpenVMS WORLD, CMKRNL, DETACH, PSWAPM, PRMMBX, ALTPRI, SYSGBL, SYSNAM, SYSPRV, and BYPASS privileges. NOTE If Oracle Trace is installed on your system, you stall the Oracle Rdb monitor process with the RMU Monitor Stop command unless you do one of the following: - Shut down Oracle Trace, then shut down the Oracle Rdb monitor (in that order). - Use the RMU Monitor Stop command with the Abort=Delprc qualifier to shut down Oracle Rdb and force the monitor out of the Oracle Trace database. 4 Examples Example 1 The following command causes the Oracle Rdb monitor process to shut down after existing database users end their access to the database. New users on this node are unable to attach to any Oracle Rdb database. $ RMU/MONITOR STOP Example 2 The following command causes the Oracle Rdb monitor to stop immediately without allowing current Oracle Rdb users to complete active transactions (they are rolled back) or detach (DISCONNECT) from their databases. However, the user processes are not deleted. Because the monitor is shut down, all Oracle Rdb activity on this node is terminated. $ RMU/MONITOR STOP /ABORT=FORCEX Example 3 The following command causes the Oracle Rdb monitor to stop immediately without allowing current Oracle Rdb users to complete active transactions (they are not rolled back) or detach (DISCONNECT) from their databases. Each user process that was attached to a Oracle Rdb database on this node is deleted immediately. $ RMU/MONITOR STOP /ABORT=DELPRC 2 Move_Area Permits you to move one or more storage areas to different disks. You can also choose to move the database root file to a different disk. 3 Description The RMU Move_Area command lets you modify certain area parameters when the move operation is performed. All the files are processed simultaneously during the move operation. The performance of the RMU Move_Area command is similar to that of the RMU Backup command, and it eliminates the need for intermediate storage media. Note that when a snapshot file is moved, Oracle RMU does not actually move the snapshot file; instead, Oracle RMU re-creates and initializes the snapshot file in the specified location. See the description of the Snapshot qualifier for more information about using this qualifier, including information on its proper usage. NOTE You must perform a full and complete Oracle RMU backup operation immediately after the Oracle RMU move area operation completes to ensure that the database can be properly restored after a database failure or corruption. 3 Format (B)0RMU/Move_Area root-file-spec storage-area-list Command Qualifiers x Defaults x /[No]After_Journal[=file-spec] x See description /[No]Aij_Options[=journal-opts-file] x See description /All_Areas x See description /[No]Area x See description /[No]Cdd_Integrate x Nocdd_Integrate /[No]Checksum_Verification x /Checksum_Verification /Directory=directory-spec x None /[No]Log x Current DCL verify value /Nodes_Max=n x Keep current value /[No]Online x Noonline /Option=file-spec x None /Page_Buffers=n x n=3 /Path=cdd-path x Existing value /[No]Quiet_Point x /Quiet_Point /Root=file-spec x None /Threads=n x /Threads=10 /Users_Max=n x Keep current value (B)0 x File or Area Qualifiers x Defaults x /Blocks_Per_Page=n x None /Extension={Disable | Enable } x Current value /File=file-spec x None /Read_Only x Current value /Read_Write x Current value /Snapshots=(Allocation=n,File=file-spec) x None /[No]Spams x Leave attribute unchanged /Thresholds=(n,n,n) x None 3 Parameters 4 root-file-spec The name of the database root file for the database whose storage areas you want to move. 4 storage-area-list The name of one or more storage areas that you want to move. 3 Command_Qualifiers 4 After_Journal After_Journal[=file-spec] Noafter_Journal NOTE This qualifier is maintained for compatibility with versions of Oracle Rdb prior to Version 6.0. You might find it more useful to specify the Aij_Options qualifier, unless you are only interested in creating extensible after-image journal (.aij) files. Specifies how Oracle RMU is to handle after-image journaling and .aij file creation, using the following rules: o If you specify the After_Journal qualifier and provide a file specification, Oracle RMU enables after-image journaling and creates a new extensible after-image journal (.aij) file for the database. o If you specify the After_Journal qualifier but do not provide a file specification, Oracle RMU enables after-image journaling and creates a new extensible .aij file for the database with the same name as, but a different version number from, the .aij file for the database root file being moved. o If you specify the Noafter_Journal qualifier, Oracle RMU disables after-image journaling and does not create a new .aij file. o If you do not specify an After_Journal, Noafter_Journal, Aij_Options, or Noaij_Options qualifier, Oracle RMU retains the original journal setting (enabled or disabled) and the original .aij file state. You can only specify one, or none, of the following after-image journal qualifiers in a single RMU Move_Area command: After_ Journal, Noafter_Journal, Aij_Options, or Noaij_Options. You cannot use the After_Journal qualifier to create fixed-size .aij files; use the Aij_Options qualifier. You can facilitate recovery by creating a new .aij file because a single .aij file cannot be applied across a move area operation that changes an area page size. A single .aij file cannot be applied across a move operation because the move operation is never recorded in the .aij file (and therefore the increase in page size is also not journaled). Therefore, when you attempt to recover the database, the original page size is used for recovery purposes. So, if the .aij file contains database insert transactions, these updates might have more free space associated with them than is available on the original page size. This results in an inability to recover the insert transaction, which in turn results in a bugcheck and a corrupted database. This qualifier is valid only when no users are attached to the database and only when the root file is moved. 4 Aij_Options Aij_Options[=journal-opts-file] Noaij_Options Specifies how Oracle RMU is to handle after-image journaling and .aij file creation, using the following rules: o If you specify the Aij_Options qualifier and provide a journal-opts-file, Oracle RMU enables journaling and creates the .aij file or files you specify for the database. If only one .aij file exists for the database, it will be an extensible .aij file. If two or more .aij files are created for the database, they will be fixed-size .aij files (as long as at least two .aij files are always available). o If you specify the Aij_Options qualifier but do not provide a journal-opts-file, Oracle RMU disables journaling and does not create any new .aij files. o If you specify the Noaij_Options qualifier, Oracle RMU retains the original journal setting (enabled or disabled) and retains the original .aij file. o If you do not specify an After_Journal, Noafter_Journal, Aij_Options, or Noaij_Options qualifier, Oracle RMU retains the original journal setting (enabled or disabled) and the original .aij file state. See Show After_Journal for information on the format of a journal-opts-file. Note that you cannot use the RMU Move_Area command with the Aij_Options qualifier to alter the journal configuration. However, you can use it to define a new after-image journal configuration. When you use it to define a new after-image journal configuration, it does not delete the journals in the original configuration. Those can still be used for recovery. If you need to alter the after-image journal configuration, you should use the RMU Set After_Journal command. The Aij_Options qualifier is valid only when no users are attached to the database and only when the root file is moved. 4 All_Areas All_Areas Noarea Specifies that all database storage areas are to be moved. If you specify the All_Areas qualifier, you do not need to specify a storage-area-list. By default, only areas specified in the storage-area-list are moved. 4 Area Area Noarea NOTE Due to the confusing semantics of the Area and Noarea qualifiers, the Area and Noarea qualifiers are deprecated. Oracle Corporation recommends that you use one of the following methods to specify areas to be moved: o To move all the storage areas in the database use the All_Areas qualifier and do not specify a storage-area- list parameter o To move only selected areas in the database, specify the storage-area-list parameter or use the Options qualifier and specify an options file. o To move only the database root file for a multifile database, or to move an entire single-file database, specify the root qualifier and do not specify a storage- area-list parameter. Controls whether specific storage areas are moved. If you specify the Area qualifier, only the storage areas specified in the option file or the storage-area-list are moved. If you specify Noarea, all the storage areas in the database are moved. The default is the Area qualifier. 4 Cdd_Integrate Cdd_Integrate Nocdd_Integrate Integrates the metadata from the root (.rdb) file of the moved database into the data dictionary (assuming the data dictionary is installed on your system). If you specify the Nocdd_Integrate qualifier, no integration occurs during the move operation. You can use the Nocdd_Integrate qualifier even if the DICTIONARY IS REQUIRED clause was used when the database being moved was defined. The Cdd_Integrate qualifier integrates definitions in one direction only-from the database file to the dictionary. The Cdd_Integrate qualifier does not integrate definitions from the dictionary to the database file. The Nocdd_Integrate qualifier is the default. 4 Checksum_Verification Checksum_Verification Nochecksum_Verification Requests that the page checksum be verified for each page moved. The default is to perform this verification. The Checksum_Verification qualifier uses CPU resources but can provide an extra measure of confidence in the quality of the data being moved. Use of the Checksum_Verification qualifier offers an additional level of data security when the database employs disk striping or RAID (redundant arrays of inexpensive disks) technology. These technologies fragment data over several disk drives, and use of the Checksum_Verification qualifier permits Oracle RMU to detect the possibility that the data it is reading from these disks has been only partially updated. If you use either of these technologies, you should use the Checksum_Verification qualifier. Oracle Corporation recommends that you use the Checksum_ Verification qualifier with all move operations where integrity of the data is essential. 4 Directory Directory=directory-spec Specifies the destination directory for the moved database files. Note that if you specify a file name or file extension, all moved files are given that file name or file extension. There is no default directory specification for this qualifier. See the Usage Notes for information on how this qualifier interacts with the Root, File, and Snapshot qualifiers and for warnings regarding moving database files into a directory owned by a resource identifier. If you do not specify this qualifier, Oracle RMU attempts to move all the database files (unless they are qualified with the Root, File, or Snapshot qualifier) to their current location. 4 Log Log Nolog Specifies whether the processing of the command is reported to SYS$OUTPUT. Specify the Log qualifier to request log output and the Nolog qualifier to prevent it. If you specify neither, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 4 Nodes_Max Nodes_Max=n Specifies a new value for the database maximum node count parameter. The default is to leave the value unchanged. Use the Nodes_Max qualifier only if you move the database root file. 4 Online Online Noonline Allows the specified storage areas to be moved without taking the database off line. This qualifier can be used only when you specify the storage-area-list parameter, or when you specify the Options=file-spec qualifier. The default is Noonline. You cannot move a database root file when the database is on line. The Root qualifier cannot be specified with the Online qualifier in an RMU Move_Area command. 4 Option Option=file-spec Specifies an options file containing storage area names, followed by the storage area qualifiers that you want applied to that storage area. Do not separate the storage area names with commas. Instead, put each storage area name on a separate line in the file. The storage area qualifiers that you can include in the options file are: Blocks_Per_Page File Snapshot Thresholds If you specify the Snapshot qualifier, you must also move the corresponding data files at the same time. To move a snapshot file independently of its corresponding data file, use the RMU Repair command with the Initialize=Snapshots=Confirm qualifier. You can use the DCL line continuation character, a hyphen (-), or the comment character (!) in the options file. There is no default for this qualifier. Example 3 in the Examples help entry under this command shows the use of an options file. If the Option qualifier is specified, the storage-area-list parameter is ignored. 4 Page_Buffers Page_Buffers=n Specifies the number of buffers to be allocated for each file to be moved. The number of buffers used is twice the number specified; half are used for reading the file and half for writing the moved files. Values specified with the Page_Buffers qualifier can range from 1 to 5. The default value is 3. Larger values might improve performance, but they increase memory usage. 4 Path Path=cdd-path Specifies a data dictionary path into which the definitions of the moved database will be integrated. If you do not specify the Path qualifier, Oracle RMU uses the CDD$DEFAULT logical name value of the user who enters the RMU Move_Area command. If you specify a relative path name, Oracle Rdb appends the relative path name you enter to the CDD$DEFAULT value. If the cdd-path parameter contains nonalphanumeric characters, you must enclose it within quotation marks (""). Oracle Rdb ignores the Path qualifier if you use the Nocdd_ Integrate qualifier or if the data dictionary is not installed on your system. 4 Quiet_Point Quiet_Point Noquiet_Point Allows you to specify that a database move operation is to occur either immediately or when a quiet point for database activity occurs. A quiet point is defined as a point where no active update transactions are in progress in the database. When you specify the Noquiet_Point qualifier, Oracle RMU proceeds with the move operation as soon as the RMU Move_Area command is issued, regardless of any update transaction activity in progress in the database. Because Oracle RMU must acquire exclusive locks on the physical and logical areas for the areas being moved, the move operation fails if there are any active transactions with exclusive locks on storage areas that are being moved. However, once Oracle RMU has successfully acquired all the needed concurrent-read storage area locks, it should not encounter any further lock conflicts. If a transaction is started that causes Oracle Rdb to request exclusive locks on the areas that are in the process of being moved, that transaction either waits or gets a lock conflict error, but the move area operation continues unaffected. If you intend to use the Noquiet_Point qualifier with a move procedure that previously specified the Quiet_Point qualifier (or did not specify either the Quiet_Point or the Noquiet_Point qualifier), you should examine any applications that execute concurrently with the move operation. You might need to modify your applications or your move procedure to handle the lock conflicts that can occur when you specify the Noquiet_Point qualifier. When you specify the Quiet_Point qualifier, the move operation begins when a quiet point is reached. The default is Quiet_Point. 4 Root Root=file-spec Requests that the database root file be moved to the specified location. If not specified, the database root file is not moved. You must specify the Root qualifier when you use the RMU Move_ Area command on a single-file database. If you omit the Root qualifier, you receive an error message. When you specify the Root qualifier, specify the location where you want the root file moved. For example: $ RMU/MOVE_AREA/ROOT=DISK1:[DATABASE.TEST] MF_PERSONNEL See the Usage Notes for information on how this qualifier interacts with the Directory, File, and Snapshot qualifiers. 4 Threads=number Threads=number Specifies the number of reader threads to be used by the move process. RMU creates so called internal 'threads' of execution to read data from one specific storage area. Threads run quasi-parallel within the process executing the RMU image. Each thread generates its own I/O load and consumes resources like virtual address space and process quotas (e.g. FILLM, BYTLM). The more threads, the more I/Os can be generated at one point in time and the more resources are needed to accomplish the same task. Performance increases with more threads due to parallel activities which keeps disk drives busier. However, at a certain number of threads, performance suffers because the disk I/O subsystem is saturated and I/O queues build up for the disk drives. Also the extra CPU time for additional thread scheduling overhead reduces the overall performance. Typically 2-5 threads per input disk drive are sufficient to drive the disk I/O susbsystem at its optimum. However, some controllers may be able to handle the I/O load of more threads, for example disk controllers with RAID sets and extra cache memory. In a move operation, one thread moves the data of one storage area at a time. If there are more storage areas to be moved than there are threads, then the next idle thread takes on the next storage area. Storage areas are moved in order of the area size - largest areas first. This optimizes the overall elapsed time by allowing other threads to move smaller areas while an earlier thread is still working on a large area. If no threads qualifier is specified then 10 threads are created by default. The minimum is 1 thread and the maximum is the number of storage areas to be moved. If the user specifies a value larger than the number of storage areas, then RMU silently limits the number of threads to the number of storage areas. For a move operation, you can specify a threads number as low as 1. Using a threads number of 1 generates the smallest system load in terms of working set usage and disk I/O load. Disk I/O subsystems most likely can handle higher I/O loads. Using a slightly larger value than 1 typically results in faster execution time. 4 Users_Max Users_Max=n Specifies a new value for the database maximum user count parameter. The default is to leave the value unchanged. Use the Users_Max qualifier only if you move the database root file. 4 Blocks_Per_Page Blocks_Per_Page=n Specifies a new page size for the storage area to which it is applied. You cannot decrease the page size of a storage area. If you attempt to change the page size during an online Move_ Area operation, you might receive a PAGESIZETOOBIG error message. Changing the page size sometimes requires that Oracle Rdb change the buffer size for the database also (because buffers must be large enough to hold at least one page from each area). However, the buffer size cannot change if other users are accessing the database. You might want to increase the page size in storage areas containing hash indexes that are close to full. By increasing the page size in such a situation, you prevent the storage area from extending. The Blocks_Per_Page qualifier is a positional qualifier. 4 Extension Extension=Disable Extension=Enable Allows you to change the automatic file extension attribute when you move a storage area. Use the Extension=Disable qualifier to disable automatic file extensions for one or more storage areas. Use the Extension=Enable qualifier to enable automatic file extensions for one or more storage areas. If you do not specify the Extension=Disable or the Extension=Enable qualifier, the storage areas is moved with the automatic file extension attributes that are currently in effect. The Extension qualifier is a positional qualifier. 4 File File=file-spec Requests that the storage area to which this qualifier is applied be moved to the specified location. The File qualifier is a positional qualifier. This qualifier is not valid for single-file databases. See the Usage Notes for information on how this qualifier interacts with the Root, Snapshot, and Directory qualifiers. 4 Read_Only Use the Read_Only qualifier to change a read/write storage area or a write-once storage area to a read-only storage area. If you do not specify the Read_Only or the Read_Write qualifier, the storage areas are moved with the read/write attributes that are currently in effect for the database. This is a positional qualifier. 4 Read_Write Use the Read_Write qualifier to change a read-only storage area or a write-once storage area to a read/write storage area. If you do not specify the Read_Only or the Read_Write qualifier, the storage areas are moved with the read/write attributes that are currently in effect for the database. This is a positional qualifier. 4 Snapshots Snapshots=(Allocation=n,File=file-spec) Allows you to specify a new snapshot file allocation size, a new snapshot file location, or both, for the storage area to which the qualifier is applied. Use the Allocation=n option to specify the snapshot file allocation size in n pages; use the File=file-spec option to specify a new file location for the snapshot file associated with the area being moved. Note that when you specify a new file location for the snapshot file, the snapshot file is not actually moved; instead, Oracle RMU creates and initializes a new snapshot file in the specified directory. However, if a snapshot file is accidentally deleted or becomes corrupt, using this qualifier is not the recommended or supported method for re-creating the snapshot file. Use the RMU Repair command instead. See the Repair help entry for information on using the RMU Repair command to re-create and initialize a deleted or corrupted snapshot file. If the keyword Allocation is omitted, the original allocation is used, not the storage area's current allocation size. You cannot specify a snapshot file name for a single-file database. When you create a snapshot file, Oracle Rdb does not store the file specification of the snapshot file. Instead, it uses the file specification of the root file (.rdb) to determine the file specification of the snapshot file. See the Usage Notes for information on placing a snapshot file on a different device or directory when your database is a single- file database and for information on how this qualifier interacts with the Root, File, and Directory qualifiers. The Snapshot qualifier is a positional qualifier. 4 Spams Spams Nospams Specifies whether to enable the creation of space area management (SPAM) pages or to disable the creation of SPAM pages (Nospams) for specified storage areas when converting read/write storage areas to write-once storage areas or vice versa. This qualifier is not permitted with a storage area that has a uniform page format. When SPAM pages are disabled in a read/write storage area, the SPAM pages are initialized, but they are not updated. The Spams qualifier is a positional qualifier. 4 Thresholds Thresholds=(n,n,n) Specifies new SPAM thresholds for the storage area to which it is applied (for a mixed page format storage area). The thresholds of a storage area with a uniform page format cannot be changed. See the Oracle Rdb7 Guide to Database Performance and Tuning for information on setting SPAM thresholds. The Thresholds qualifier is a positional qualifier. 3 Usage_Notes o To use the RMU Move_Area command for a database, you must have the RMU$MOVE privilege in the root file access control list (ACL) for the database or have the OpenVMS SYSPRV or BYPASS privilege. o You cannot disable extensions of snapshot (.snp) files. o The parameter (file and area) qualifiers for the RMU Move_Area command have positional semantics. See the Command_Qualifiers help entry for more information on parameter qualifiers. o The RMU Move_Area command provides four qualifiers, Directory, Root, File, and Snapshots, that allow you to specify the target for the moved files. The target can be just a directory, just a file name, or a directory and file name. If you use all or some of these four qualifiers, apply them as follows: - If you want to move the database root, use the Root qualifier to indicate the target for the moved database root file. - Use local application of the File qualifier to specify the target for the moved storage area or areas. - Use local application of the Snapshots qualifier to specify the target for the moved snapshot file or files. - Use the Directory qualifier to specify a default target directory. The default target directory is the directory to which all storage area and snapshot files not qualified with the File or Snapshot qualifier are moved. It is also the default directory for files qualified with the Root, File, or Snapshot qualifier if the target for these qualifiers does not include a directory specification. Note the following when using these qualifiers: - Global application of the File qualifier when the target specification includes a file name causes Oracle RMU to move all of the specified storage areas to different versions of the same file name. This creates a database that is difficult to manage. - Global application of the Snapshot qualifier when the target specification includes a file name causes Oracle RMU to move all of the specified snapshot files to different versions of the same file name. This creates a database that is difficult to manage. - Specifying a file name or extension with the Directory qualifier is permitted, but causes Oracle RMU to move all of the specified files (except those specified with the File or Root qualifier) to different versions of the same file name. Again, this creates a database that is difficult to manage. See Example 6. o You must specify the Root qualifier when you use the RMU Move_ Area command on a single-file database. If you omit the Root qualifier, you receive an error message. If you want to place the snapshot file for a single-file database on a different device or directory from the root file, Oracle Corporation recommends that you create a multifile database. However, you can work around this restriction by defining a search list for a concealed logical name. (However, do not use a nonconcealed rooted logical name to define database files; a database created with a non-concealed rooted logical name can be backed up, but may not restore correctly when you attempt to restore the files to a new directory.) To create a single-file database with a snapshot file on a different device or directory from the root file, define a search list by using a concealed logical name. Specify the location of the root file as the first item in the search list. When you create the database, use the logical name for the directory specification. Then, copy the snapshot file to the second device. The following example demonstrates the workaround: $ ! Define a concealed logical name. $ DEFINE /TRANS=CONCEALED/SYSTEM TESTDB USER$DISK1:[DATABASE], - _$ USER$DISK2:[SNAPSHOT] $ $ SQL SQL> ! Create the database. SQL> ! SQL> CREATE DATABASE FILENAME TESTDB:TEST; SQL> EXIT $ ! $ ! Copy the snapshot file to the second disk. $ COPY USER$DISK1:[DATABASE]TEST.SNP USER$DISK2:[SNAPSHOT]TEST.SNP $ ! $ ! Delete the snapshot file from the original disk. $ DELETE USER$DISK1:[DATABASE]TEST.SNP; o There are no restrictions on the use of the Nospams qualifier option with mixed page format storage areas, but the use of the Nospams qualifier typically causes severe performance degradation. The Nospams qualifier is useful only where updates are rare and batched, and access is primarily by database key (dbkey). 3 Examples Example 1 If a storage area is on a disk that is logging error messages, you can move the storage area to another disk by using the RMU Move_Area command. The following command moves the DEPARTMENTS storage area (departments.rda) and the DEPARTMENTS snapshot file (departments.snp) of the mf_personnel database to the DDV21:[RICK.SQL] directory: $ RMU/MOVE_AREA MF_PERSONNEL DEPARTMENTS /DIRECTORY=DDV21:[RICK.SQL] Example 2 The following command moves the EMPIDS_LOW, EMPIDS_MID, and EMPIDS_OVER storage areas for the mf_personnel database to the DISK2:[USER2] directory. The Extension=Disable qualifier disables automatic file extensions for the EMPIDS_LOW, EMPIDS_MID, and EMPIDS_OVER storage area (.rda) files when they are moved to the DISK2:[USER2] directory: $ RMU/MOVE_AREA/EXTENSION=DISABLE/DIRECTORY=DISK2:[USER2] - _$ mf_personnel EMPIDS_LOW,EMPIDS_MID,EMPIDS_OVER Example 3 The following RMU Move_Area command uses an options file to specify that the storage area files and snapshot files be moved to different disks. Note that storage area snapshot (.snp) files are located on different disks from one another and from their associated storage area (.rda) files; this is recommended for optimal performance. (This example assumes that the disks specified for each storage area file in options_file.opt are different from those where the storage area files currently reside.) $ RMU/MOVE_AREA/OPTIONS=OPTIONS_FILE.OPT MF_PERSONNEL The following command displays the contents of the options file: $ TYPE options_file.opt EMPIDS_LOW /FILE=DISK1:[CORPORATE.PERSONNEL]EMPIDS_LOW.RDA - /SNAPSHOT=(FILE=DISK2:[CORPORATE.PERSONNEL]EMPIDS_LOW.SNP) EMPIDS_MID /FILE=DISK3:[CORPORATE.PERSONNEL]EMPIDS_MID.RDA - /SNAPSHOT=(FILE=DISK4:[CORPORATE.PERSONNEL]EMPIDS_MID.SNP) EMPIDS_OVER /FILE=DISK5:[CORPORATE.PERSONNEL]EMPIDS_OVER.RDA - /SNAPSHOT=(FILE=DISK6:[CORPORATE.PERSONNEL]EMPIDS_OVER.SNP) DEPARTMENTS /FILE=DISK7:[CORPORATE.PERSONNEL]DEPARTMENTS.RDA - /SNAPSHOT=(FILE=DISK8:[CORPORATE.PERSONNEL]DEPARTMENTS.SNP) SALARY_HISTORY /FILE=DISK9:[CORPORATE.PERSONNEL]SALARY_HISTORY.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]SALARY_HISTORY.SNP) JOBS /FILE=DISK7:[CORPORATE.PERSONNEL]JOBS.RDA - /SNAPSHOT=(FILE=DISK8:[CORPORATE.PERSONNEL]JOBS.SNP) EMP_INFO /FILE=DISK9:[CORPORATE.PERSONNEL]EMP_INFO.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]EMP_INFO.SNP) RESUME_LISTS /FILE=DISK11:[CORPORATE.PERSONNEL]RESUME_LISTS.RDA - /SNAPSHOT=(FILE=DISK12:[CORPORATE.PERSONNEL]RESUME_LISTS.SNP) RESUMES /FILE=DISK9:[CORPORATE.PERSONNEL]RESUMES.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]RESUMES.SNP) Example 4 The following RMU Move_Area command moves the database root for the mf_personnel database and defines a new after-image journal configuration, using the Aij_Options qualifier: $ RMU/MOVE_AREA/ROOT=DISK1:[DATABASE.PERSONNEL]MF_PERSONNEL - _$ /AIJ_OPTIONS=aij_config.opt MF_PERSONNEL/NOONLINE The aij_config.opt file contains the following clauses: JOURNAL IS ENABLED - RESERVE 2 - ALLOCATION IS 512 - EXTENT IS 512 - OVERWRITE IS DISABLED - SHUTDOWN_TIMEOUT IS 120 - NOTIFY IS DISABLED - BACKUPS ARE MANUAL - CACHE IS DISABLED ADD AIJ1 - FILE DISK2:[MFPERS_AIJ1]AIJ_ONE ADD AIJ2 - FILE DISK3:[MFPERS_AIJ2]AIJ_TWO Example 5 The following example moves all the mf_personnel database storage areas to the DISK3:[db] directory: $ RMU/MOVE_AREA MF_PERSONNEL.RDB /ALL_AREAS/DIR=DISK3:[DB] Example 6 The following example demonstrates the use of the Directory, File, and Root qualifiers. In this example: o The default directory is specified as DISK2:[DIR]. o The target directory and file name for the database root file is specified with the Root qualifier. The target directory specified with the Root qualifier overrides the default directory specified with the Directory qualifier. Thus, Oracle RMU moves the database root to DISK3:[ROOT] and names it MOVEDRDB.RDB. o The target directory for the EMPIDS_MID storage area is DISK4:[FILE]. Oracle RMU moves EMPIDS_MID to DISK4:[FILE]. o The target file name for the EMPIDS_LOW storage area is EMPIDS. Thus, Oracle RMU moves the EMPIDS_LOW storage area to the DISK2 default directory (specified with the Directory qualifier), and names the file EMPIDS.RDA. o The target for the EMPIDS_LOW snapshot file is DISK5:[SNAP]EMPIDS.SNP Thus, Oracle RMU moves the EMPIDS_LOW snapshot file to DISK5:[SNAP]EMPIDS.SNP. o All the other storage area files and snapshot files in the mf_personnel database are moved to DISK2:[DIR]; the file names for these storage areas remain unchanged. $ RMU/MOVE_AREA DISK1:[DB]MF_PERSONNEL.RDB /ALL- _$ /DIRECTORY=DISK2:[DIR] - _$ /ROOT=DISK3:[ROOT]MOVEDRDB.RDB - _$ EMPIDS_MID/FILE=DISK4:[FILE], - _$ EMPIDS_LOW/FILE=EMPIDS - _$ /SNAPSHOT=(FILE=DISK5:[SNAP]EMPIDS.SNP) 2 Open Opens a database root file and maps its global section to the contents of an OpenVMS virtual address file. You can use the RMU Open command in conjunction with the SQL ALTER DATABASE statement to control access to the database. See the description of the OPEN IS {AUTOMATIC | MANUAL} clause of the SQL ALTER DATABASE statement in the Oracle Rdb SQL Reference Manual for details. 3 Description Once you use the RMU Open command to open a database, the database remains open and mapped until you close it explicitly with an RMU Close command and all users have exited the database with the SQL DISCONNECT or EXIT statements. If you do not issue the RMU Open command, the first user to attach to the database incurs the cost of implicitly opening it and the last user to detach from the database incurs the cost of implicitly closing it. The effect of the RMU Open command depends on whether you have specified the OPEN IS AUTOMATIC or OPEN IS MANUAL clause to the SQL ALTER DATABASE statement, as follows: o OPEN IS AUTOMATIC If you have specified automatic opening for your database, users can invoke the database at any time without first issuing an RMU Open command. (Although as mentioned above, it is more efficient to explicitly open the database with an RMU Open command and close it with an RMU Close command.) o OPEN IS MANUAL If you have specified manual opening for your database, the RMU Open command must be issued before users can invoke the database. If you modify the database attribute from OPEN IS AUTOMATIC to OPEN IS MANUAL, the modification takes effect only after all users have detached from the database. (You can issue the RMU/CLOSE/ABORT=FORCEX command to force all users to detach.) Then, you must issue the RMU Open command before users can invoke the database. If you modify the database attribute from OPEN IS MANUAL to OPEN IS AUTOMATIC, users can invoke the database at their discretion. You do not have to issue the RMU Open command. However, if a user has already opened the database manually when you make this change to the database attribute, the modification takes effect only after you manually close the database by issuing the RMU Close command. See the Oracle Rdb Guide to Database Maintenance for information to help you decide whether to set your database attribute to automatic or manual opening. When you create a database, you have a choice of how to set up buffers for database pages. You can choose either local or global buffering. Global buffers can provide better system performance. See the Oracle Rdb7 Guide to Database Performance and Tuning for more information on setting the number of global buffers for your system. 3 Format (B)0 Command Qualifiers x Default x /Access=[Un]Restricted x See description /Global_Buffers[=(Total=i,User_Limit=j)] x See description /Path x None /Row_Cache=Disable x See description /[No]Statistics=Import x /Nostatistics /[No]Wait x /Nowait 3 Parameters 4 root-file-spec root-file-spec[,...] Specifies the database to open. If the database root file is open, you receive an informational message. The default file extension is .rdb. 3 Command_Qualifiers 4 Access Access=Restricted Access=Unrestricted Permits the database administrator to open the database and restrict access to it in order to perform maintenance operations or to restructure the database without interference from users who want to gain access. If access is restricted (Access=Restricted), the DBADM privilege is required for SQL access to the database. If the Access=Unrestricted qualifier is specified, users without the DBADM privilege can attach to the database. NOTE Do not confuse the Oracle RMU Access=Restricted qualifier with the SQL restricted access clause (available for use with the following SQL statements: ATTACH, CREATE, DECLARE ALIAS, and IMPORT). When you specify the restricted access clause in SQL, only one user can attach to the database; when you specify the Access=Restricted qualifier using Oracle RMU, any number of users with the DBADM privilege can access the database. Furthermore, note that an SQL SHOW DATABASE command displays the phrase "No Restricted Access" or the phrase "Restricted Access" if access has been restricted using the SQL restricted access clause. However, SHOW DATABASE tells you nothing about whether Oracle RMU has opened a database with access restricted. Use the RMU Dump command to view the Oracle RMU access setting. Refer to the Oracle Rdb SQL Reference Manual for more information on the SQL restricted access clause. If you specify the RMU Open command without the Access qualifier, Oracle RMU opens the database in the same access mode as the last RMU Open command performed. If the database was last opened as restricted, issuing the RMU Dump command results in the following message being displayed: Access restricted to privileged users Use this form of the RMU Open command to open the database on other nodes without changing the access mode. The access mode is clusterwide and the last mode set with the RMU Open command is used for the entire cluster. For example, if you open the mf_personnel database on node A with the Access=Unrestricted qualifier, and open the same database on node B with the Access=Restricted qualifier, the database has restricted access on both node A and node B. However, the commands do not terminate any user processes that may have gained access while the database was unrestricted. The access mode is stored in the database. Consequently, if the system fails while access is restricted, access remains restricted unless the unrestricted mode is explicitly requested. The RMU Backup, RMU Restore, and RMU Copy_Database commands also preserve the access mode. The RMU Close command does not alter the access mode. You can change the mode by using the RMU Open command only. You can use the RMU Open command to restrict access to any database, whether it was opened as AUTOMATIC or MANUAL. The Access qualifier is a positional qualifier. 4 Global_Buffers Global_Buffers[=(Total=i,User_Limit=j)] Allows you to set the basic global buffer parameters on each RMU Open command. If you specify the Global_Buffers qualifier, you can optionally specify values for the Total and User_Limit parameters: o Total is the number of global buffers per node to allocate for this opened instance of the database (minimum = 5, and maximum = 500,000). o User_Limit is the maximum number of global buffers to be allotted to any given user (minimum = 5, maximum = Total). The default values for Total and User_Limit are set by: o The RMU Open command explicitly o Values determined at the time the database was created If you do not specify a value for the Total or User_Limit options, the values are determined based on what they were when the database was created. If a database does not have global buffers enabled, the Global_ Buffers qualifier is ignored. Use the RMU Dump command to see if global buffering is enabled or disabled. The RMU Dump command also shows the global buffer count and the maximum global buffer count per user. For example: $ RMU/DUMP MF_PERSONNEL *------------------------------------------------------------------ * Oracle Rdb V7.0-00 22-SEP-1995 10:11:51.14 * * Dump of Database header * Database: DISK1:[DATABASE]MF_PERSONNEL.RDB;1 * *------------------------------------------------------------------- Database Parameters: Root filename is "DISK1:[DATABASE]MF_PERSONNEL.RDB;1" Created at 7-APR-1994 16:50:09.01 Oracle Rdb structure level is 70.0 Maximum user count is 50 Maximum node count is 16 Database open mode is AUTOMATIC Database close mode is AUTOMATIC Database is available for READ WRITE access Snapshot mode is NON-DEFERRED Statistics are enabled Storage Areas... - Active storage area count is 10 - Reserved storage area count is 0 Buffers... - Default user buffer count is 20 - Default recovery buffer count is 20 - Global buffers are enabled <-------- - Global buffer count is 250 <-------- - Maximum global buffer count per user is 5 <-------- - Buffer size is 6 blocks . . . Derived Data... - Global section size With global buffers disabled is 70962 bytes With global buffers enabled is 975992 bytes . . . The Global_Buffers qualifier is a positional qualifier. 4 Path Path Specifies the full or relative data dictionary path name in which the definitions reside for the database you want to open. The Path qualifier is a positional qualifier. The path name cannot include wildcard characters. 4 Row_Cache=Disable Disables row caching. This qualifier is provided for use with hot standby databases. Row caching cannot be enabled on a hot standby database while replication is active. If it is enabled, the hot standby feature will not start. 4 Statistics=Import Statistics=Import Nostatistics Specifies that statistic information previously saved by using the Statistics=Export qualifier on the RMU Close command is to be loaded when the database is opened. The default is Nostatistics, which indicates that statistic information is not loaded when the database is opened. After the database is opened using the Statistics=Import qualifier, the saved statistics file is closed. The statistics file is not automatically deleted. It can be deleted if it is no longer needed. When you use the Statistics=Import qualifier, statistics information is automatically preserved in the event of abnormal database closure. To ensure that the ondisk statistic information files are accurate in the case of a node or monitor failure, the statistic information files are checkpointed by the database monitor every half-hour. The RMU Show Users command identifies when the checkpoint for each database occurs. The statistic files are not loaded if the physical schema of the database has changed since the statistic file was created. This means that the addition or deletion of storage aras, logical areas, and record caches invalidate the statistic files. This restriction prevents incorrect statistic information from being loaded when intervening physical changes occur to the database. Closing the database updates the statistic files and makes them valid. Use the RMU Show Users command to verify that the statistic information file was imported. 4 Wait Wait Nowait Specifies whether the system prompt should be returned before the database is completely open and available. Specify the Wait qualifier if you want the system prompt returned when the database is completely open and available. Specify Nowait if you want the system prompt returned immediately, regardless of the state of the open operation. The Nowait qualifier is the default. 3 Usage_Notes o To use the RMU Open command for a database, you must have the RMU$OPEN privilege in the root file access control list (ACL) for the database or the OpenVMS WORLD privilege. 3 Examples Example 1 The following command opens the mf_personnel database: $ RMU/OPEN MF_PERSONNEL Example 2 The following command opens the mf_personnel database in the WORK directory, all the databases in the .TEST directory, and the databases specified by the path names CDD$TOP.FINANCE and SAMPLE_ DB: $ RMU/OPEN DISK1:[WORK]MF_PERSONNEL, CDD$TOP.FINANCE/PATH, - _$ DISK1:[TEST]*, SAMPLE_DB/PATH Example 3 This command opens the mf_personnel database, sets the total global buffers for this opened instance of the database, and sets the maximum number of global buffers that can be given to any user. This example limits the number of users who can access this database at any given time to 2 (Total divided by User_Limit). You may want to increase the values of Total and User_Limit. $ RMU/OPEN MF_PERSONNEL/GLOBAL_BUFFERS=(TOTAL=10,USER_LIMIT=5) If you define a user limit value that is greater than the value you specify for Total, you receive an error message: $ RMU/OPEN MF_PERSONNEL/GLOBAL=(TOTAL=5,USER_LIMIT=10) %RMU-F-VALGTRMAX, value (10) is greater than maximum allowed value (5) for GLOBAL_BUFFERS.USER_LIMIT Example 4 This command disables row caching. $ RMU/OPEN MF_PERSONNEL.RDB /ROW_CACHE=DISABLE 2 Optimize Optimize After_Journal Optimizes a backed up after-image journal (.aij) file for database recovery (rollforward) operations by eliminating unneeded and duplicate journal records, and by ordering journal records. An optimized .aij (.oaij) file created by the RMU Optimize After_Journal command provides better recovery performance for your database than an .aij file. A benefit of this improved recovery performance is that the database is made available to users sooner. The RMU Optimize After_Journal command is used to read a backed up .aij file on disk and write the .oaij file to tape or disk. 3 Description The RMU Optimize After_Journal command performs the following optimizations to backed up .aij files: o The .aij records from transactions that rolled back are eliminated. Because transactions that are rolled back in an .aij file are not needed in a recovery operation, they are not part of an optimized .aij file. o Duplicate .aij records are eliminated. Duplicate .aij records are .aij records that update the same database record. During the rollforward of an .aij file, duplicate .aij records cause a database record to be updated multiple times. Each update supersedes the previous update, meaning only the last update is relevant. Therefore, all but the last update to a database record can be eliminated from an .aij file. o The .aij records are ordered by physical database key (dbkey). Ordering .aij records by physical dbkey improves I/O performance at recovery time. See the Oracle Rdb Guide to Database Maintenance for further description of optimizing .aij files. The RMU Optimize After_Journal command has the following restrictions: o You can only optimize quiet-point .aij backup files. o You cannot optimize a current .aij file. o You cannot optimize an .oaij file. NOTE Because an .oaij file is not functionally equivalent to the original .aij file, the original .aij file should not be discarded after it has been optimized. o You cannot use .oaij files with the following types of recovery operations: - By-area recovery operations (recovery operations that use the RMU Recover command with the Areas qualifier). - By-page recovery operations (recovery operations that use the RMU Recover command with the Just_Corrupt qualifier). - RMU Recover commands with the Until qualifier. The .oaij file does not retain enough of the information from the original .aij file for such an operation. - Recovery operation where the database or any storage areas (or both) are inconsistent with the .oaij file. A database or storage area will be inconsistent with the .oaij file if the transaction sequence number (TSN) of the last committed transaction of the database or storage area is not equal to the TSN of the last committed transaction in the open record of the .aij file. The last committed TSN in the .oaij file represents the last transaction committed to the database at the time the original .aij file was created. As a workaround for these restrictions against using .oaij files in these recovery operations, use the original, unoptimized .aij files in these recovery operations instead. o Any .aij file that possibly contains incomplete transactions cannot be optimized. Incomplete transactions can occur in an .aij file under the following circumstances: - The .aij file is backed up with a no-quiet-point backup operation (because transactions can span .aij files) Note that transactions in a fixed-size journal configuration may span .aij files. Thus, if each journal in a fixed-size journal configuration has been backed up on a per-journal basis, the resulting files are equivalent to a no-quiet-point .aij backup operation. These .aij backup files cannot be optimized unless you perform a manual quiet-point backup operation first. A quiet-point backup operation forces a switch-over to another available .aij file which ensures that no transaction spans two journal files. - The previous .aij file was backed up with a no-quiet-point backup operation - The .aij file has unresolved distributed transactions There are no workarounds to these restrictions against optimizing .aij files with incomplete transactions. 3 Format (B)0RMU/Optimize/After_Journal aij-file optimized-aij-file Command Qualifiers x Defaults x /[No]Accept_Label x /Noaccept_Label /Active_IO=max-writes x /Active_IO=3 /Block_Size=integer x See description /Crc[=Autodin_II] x See description /Crc=Checksum x See description /Nocrc x See description /Density=density-value[,[No]Compaction]x See description /Encrypt=({Value=|Name=}[,Algorithm=]) x See description /Format={Old_File|New_Tape} x /Format=Old_File /[No]Group_Size=interval x See description /Label=(label-name-list) x See description /Librarian[=options] x None /[No]Log x Current DCL verify value /[No]Media_Loader x See description /Owner=user-id x See description /Protection[=openvms-file-protection] x See description 3 Parameters 4 aij-file The name of the .aij file that you want to optimize. It cannot be a current .aij file. The default file extension is .aij. 4 optimized-aij-file The name of the optimized .oaij file to be produced by the RMU Optimize After_Journal command. The default file extension is .oaij. 3 Command_Qualifiers 4 Accept_Label Accept_Label Specifies that Oracle RMU should keep the current tape label it finds on a tape during an optimize-to-tape operation even if that label does not match the default label or that specified with the Label qualifier. Operator notification does not occur unless the tape's protection, owner, or expiration date prohibit writing to the tape. However, a message is logged (assuming logging is enabled) to indicate that a label is being preserved and which drive currently holds that tape. This qualifier is particularly useful when your optimize-to-tape operation employs numerous previously used (and thus labeled) tapes and you want to preserve the labels currently on the tapes. If you do not specify this qualifier, the default behavior of Oracle RMU is to notify the operator each time it finds a mismatch between the current label on the tape and the default label (or the label you specify with the Label qualifier). See the description of the Labels qualifier under this command for information on default labels. See How Tapes are Relabeled During a Backup Operation in the Usage_Notes help entry under the Backup Database help entry for a summary of which labels are applied under a variety of circumstances. 4 Active_IO Active_IO=max-writes Specifies the maximum number of write operations to the .oaij file device that the RMU Optimize After_Journal command will attempt simultaneously. This is not the maximum number of write operations in progress; that value is the product of active system I/O operations and the number of devices being written to simultaneously. The value of the Active_IO qualifier can range from 1 to 5. The default value is 3. Values larger than 3 might improve performance with some tape drives. 4 Block_Size Block_Size=integer Specifies the maximum record size for the optimized .oaij file. The size can vary between 2048 and 65,024 bytes. The default value is device dependent. The appropriate block size is a compromise between tape capacity and error rate. The block size you specify must be larger than the largest page length in the database. 4 Crc[=Autodin_II] Crc[=Autodin_II] Uses the AUTODIN-II polynomial for the 32-bit cyclic redundancy check (CRC) calculation and provides the most reliable end- to-end error detection. This is the default for NRZ/PE (800/1600 bits/inch) tape drives. Typing the Crc qualifier is sufficient to select the Crc=Autodin_ II option. It is not necessary to type the entire qualifier. 4 Crc=Checksum Crc=Checksum Uses one's complement addition, which is the same computation used to do a checksum of the AIJ data on disk. This is the default for GCR (6250 bits/inch) tape drives and for TA78, TA79, and TA81 drives. The Crc=Checksum qualifier allows detection of data errors. 4 Nocrc Nocrc Disables end-to-end error detection. This is the default for TA90 (IBM 3480 class) drives. NOTE The overall effect of the Crc=Autodin_II, Crc=Checksum, and Nocrc defaults is to make tape reliability equal to that of a disk. If you retain your tapes longer than 1 year, the Nocrc default might not be adequate. For tapes retained longer than 1 year, use the Crc=Checksum qualifier. If you retain your tapes longer than 3 years, you should always use the Crc=Autodin_II qualifier. Tapes retained longer than 5 years could be deteriorating and should be copied to fresh media. See the Oracle Rdb Guide to Database Maintenance for details on using the Crc qualifiers to avoid underrun errors. 4 Density Density=density-value[,[No]Compaction] Specifies the density at which the output volume is to be written. The default value is the format of the first volume (the first tape you mount). You do not need to specify this qualifier unless your tape drives support data compression or more than one recording density. The Density qualifier is applicable only to tape drives. Oracle RMU returns an error message if this qualifier is used and the target device is not a tape drive. If your systems are running OpenVMS versions prior to 7.2-1, specify the Density qualifier as follows: o For TA90E, TA91, and TA92 tape drives, specify the number in bits per inch as follows: - Density = 70000 to initialize and write tapes in the compacted format - Density = 39872 or Density = 40000 for the noncompacted format o For SCSI (Small Computer System Interface) tape drives, specify Density = 1 to initialize and write tapes, using the drive's hardware data compression scheme. o For other types of tape drives, you can specify a supported Density value between 800 and 160,000 bits per inch. o For all tape drives, specify Density = 0 to initialize and write tapes at the drive's standard density. Do not use the Compaction or NoCompaction keyword for systems running OpenVMS versions prior to 7.2-1. On these systems, compression is determined by the density value and cannot be specified. Oracle RMU supports the OpenVMS tape density and compression values introduced in OpenVMS Version 7.2-1. The following table lists the added density values supported by Oracle RMU. DEFAULT 800 833 1600 6250 3480 3490E TK50 TK70 TK85 TK86 TK87 TK88 TK89 QIC 8200 8500 8900 DLT8000 SDLT SDLT320 SDLT600 DDS1 DDS2 DDS3 DDS4 AIT1 AIT2 AIT3 AIT4 LTO2 LTO3 COMPACTION NOCOMPACTION If the OpenVMS Version 7.2-1 density values and the previous density values are the same (for example, 800, 833, 1600, 6250), the specified value is interpreted as an OpenVMS Version 7.2-1 value if the tape device driver accepts them, and as a previous value if the tape device driver accepts previous values only. For the OpenVMS Version 7.2-1 values that accept tape compression you can use the following syntax: /DENSITY = (new_density_value,[No]Compaction) In order to use the Compaction or NoCompaction keyword, you must use one of the following density values that accepts compression: DEFAULT 3480 3490E 8200 8500 8900 TK87 TK88 TK89 DLT8000 SDLT SDLT320 AIT1 AIT2 AIT3 AIT4 DDS1 DDS2 DDS3 DDS4 SDLT600 LTO2 LTO3 Refer to the OpenVMS documentation for more information about density values. 4 Encrypt Encrypt=({Value=|Name=}[,Algorithm=]) The Encrypt qualifier decrypts the backup file of the optimized after-image journal file. Specify a key value as a string or, the name of a predefined key. If no algorithm name is specified the default is DESCBC. For details on the Value, Name and Algorithm parameters see HELP ENCRYPT. This feature requires the OpenVMS Encrypt product to be installed and licensed on this system. This feature only works for a newer format backup file which has been created using the Format=New_Tape qualifier. Therefore you have to specify the Format=New_Tape qualifier with this command if you also use the Encrypt qualifier. Synonymous with the Format=Old_File and Format=New_Tape qualifiers. See the description of those qualifiers. 4 Format Format=Old_File Format=New_Tape The Format qualifier allows you to specify the format of the files written by the RMU Optimize After_Journal command. If you specify the default, Format=Old_File, the RMU Optimize After_Journal command writes files in RMS format. This format is provided for compatibility with prior versions of Oracle Rdb. If you specify Format=Old_File, you must mount the media by using the DCL MOUNT command before you issue the RMU Optimize After_ Journal command. Because the RMU Optimize After_Journal command will use RMS to write to the tape, the tape must be mounted as an OpenVMS volume (that is, do not specify the /FOREIGN qualifier with the MOUNT command). If you specify FOREIGN access although your backup file was created using the Format=Old_File qualifier, you will not receive an error message. The tape will be considered unlabeled, and thus the operation will process whatever data is at the current position of the tape (labels, data, or something else). A failure will occur, but what will fail and how it will fail is unpredictable because the type of information that will be read is unknown. The result is an unlabeled tape that can be difficult to use for recovery operations. If you specify Format=New_Tape, the RMU Optimize After_Journal command writes .aij files in a format similar to that used by an RMU Backup command. If you specify Format=New_Tape, you must mount the media by using the DCL MOUNT command before you issue the RMU Optimize After_Journal command. The tape must be mounted as a FOREIGN volume. The following tape qualifiers have meaning only when used in conjunction with the Format=New_Tape qualifier: Active_IO Block_Size Crc Group_Size Density Label Owner_Uic Protection Rewind Tape_Expiration Follow these steps when you optimize an .aij file to tape: 1. Use the RMU Backup After_Journal command with the Format=Old_ File qualifier to back up the .aij file to disk. 2. Use the RMU Optimize After_Journal command with the Format=New_Tape qualifier to optimize the backed up .aij file to tape. 3. Use the DCL BACKUP command to create a copy of the backed up .aij file as insurance. If you enter the RMU Optimize After_Journal command with no Format qualifier, the default is Format=Old_File. 4 Group_Size Group_Size=interval Nogroup_Size Specifies the frequency at which XOR recovery blocks are written to tape. The group size can vary from 0 to 100. Specifying a group size of zero or specifying the Nogroup_Size qualifier results in no XOR recovery blocks being written. The Group_Size qualifier is applicable only to tape, and its default value is device dependent. Oracle RMU returns an error message if this qualifier is used and the target device is not a tape device. 4 Label Label=(label-name-list) Specifies the 1- to 6-character string with which the volumes of the .oaij file are to be labeled. The Label qualifier is applicable only to tape volumes. You must specify one or more label names when you use the Label qualifier. You can specify a list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas, and enclose the list of names within parentheses. Use the label that you specify for the RMU Optimize After_Journal command when you issue the RMU Recover command. The Label qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 4 Librarian Librarian[=options] Use the Librarian qualifier to back up files to data archiving software applications that support the Oracle Media Management interface. The backup file name specified on the command line identifies the stream of data to be stored in the Librarian utility. If you supply a device specification or a version number it will be ignored. The Librarian qualifier accepts the following options: o Trace_file=file-specification The Librarian utility writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian utility. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian utility. The pre-determined values are : - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical_name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian utility can use to specify catalogs or archives where Oracle Rdb backup files are stored, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of any Oracle RMU command that accesses the Librarian utility. The following OpenVMS logical names must be defined for use with a Librarian utility before you execute an Oracle RMU backup or restore operation. Do not use the Logical_Names option provided with the Librarian qualifier to define these logical names. o RMU$LIBRARIAN_PATH This logical name must be defined so that the shareable Librarian image can be loaded and called by Oracle RMU backup and restore operations. The translation must include the file type (for example, .exe), and must not include a version number. The shareable Librarian image must be an installed (known) image. See the Librarian utility documentation for the name and location of this image and how it should be installed. o RMU$DEBUG_SBT This logical name is not required. If it is defined, Oracle RMU will display debug tracing information messages from modules that make calls to the Librarian shareable image. You cannot use device specific qualifiers such as Rewind, Density, or Label with the Librarian qualifier because the Librarian utility handles the storage meda, not Oracle RMU. 4 Log Log Nolog Specifies that the optimization of the .aij file be reported to SYS$OUTPUT. When optimization activity is logged, the output from the Log qualifier provides the number of transactions committed and rolled back. You can specify the Trace qualifier with the Log qualifier. The default is the setting of the DCL VERIFY flag, which is controlled by the DCL SET VERIFY command. 4 Media_Loader Media_Loader Nomedia_Loader Use the Media_Loader qualifier to specify that the tape device receiving the backup file has a loader or stacker. Use the Nomedia_Loader qualifier to specify that the tape device does not have a loader or stacker. By default, if a tape device has a loader or stacker, Oracle RMU should recognize this fact. However, occasionally Oracle RMU does not recognize that a tape device has a loader or stacker. Therefore, when the first backup tape fills, Oracle RMU issues a request to the operator for the next tape, instead of requesting the next tape from the loader or stacker. Similarly, sometimes Oracle RMU behaves as though a tape device has a loader or stacker when actually it does not. If you find that Oracle RMU is not recognizing that your tape device has a loader or stacker, specify the Media_Loader qualifier. If you find that Oracle RMU expects a loader or stacker when it should not, specify the Nomedia_Loader qualifier. Synonymous with the Owner qualifier. See the description of the Owner qualifier. 4 Owner Owner=user-id Specifies the owner of the tape volume set. The owner is the user who will be permitted to recover (roll forward) the database. The user-id parameter must be one of the following types of OpenVMS identifier: o A user identification code (UIC) in [group-name,member-name] alphanumeric format o A UIC in [group-number,member-number] numeric format o A general identifier, such as SECRETARIES o A system-defined identifier, such as DIALUP When used with tapes, the Owner qualifier applies to all continuation volumes. The Owner qualifier applies to the first volume only if the Rewind qualifier is also specified. If the Rewind qualifier is not specified, the optimization operation appends the file to a previously labeled tape, so the first volume can have a different protection than the continuation volumes. 4 Protection Protection[=openvms-file-protection] Specifies the system file protection for the .oaij file produced by the RMU Optimize After_Journal command. The default file protection varies, depending on whether you write the .oaij file to disk or tape. This is because tapes do not allow delete or execute access and the SYSTEM account always has both read and write access to tapes. In addition, a more restrictive class accumulates the access rights of the less restrictive classes. If you do not specify the Protection qualifier, the default protection is as follows: o S:RWED,O:RE,G,W if the .oaij file is written to disk o S:RW,O:R,G,W if the .oaij file is written to tape If you specify the Protection qualifier explicitly, the differences in protection applied for backups to tape or disk as noted in the preceding paragraph are applied. Thus, if you specify Protection=(S,O,G:W,W:R), that protection on tape becomes (S:RW,O:RW,G:RW,W:R). 4 Recovery_Method Recovery_Method=Sequential Recovery_Method=Scatter Specifies how .aij records are to be ordered. You can specify one of two possible order types: o Sequential-.aij records are ordered by physical dbkey in an area:page:line sequence. This order type is the default. o Scatter-.aij records are ordered by a sort key of page:area:line (page number, area number, and line number). This order can allow the RMU Recover command to perform more effective I/O prefetching and writing to multiple storage areas simultaneously, typically where storage areas of the database are distributed among multiple disk devices. Scatter ordering allows more disk devices to be active during the recovery process. This helps reduce idle CPU time and allows the recovery to complete in less time. However, because database configurations vary widely, Oracle recommends that you perform tests with both Scatter and Sequential ordering of the optimized after-image journals to determine which method produces the best results for your system. 4 Rewind Rewind Norewind Specifies that the tape that will contain the .oaij file be rewound before processing begins. The tape will be initialized according to the Label qualifier. The Norewind qualifier is the default and causes the optimized .oaij file to be written starting at the current logical end-of-tape (EOT). The Norewind and Rewind qualifiers are applicable only to tape devices. Oracle RMU returns an error message if these qualifiers are used and the target device is not a tape device. 4 Tape_Expiration Tape_Expiration=date-time Specifies the expiration date of the .oaij file on tape. Note that when Oracle RMU reads a tape, it looks at the expiration date in the file header of the first file on the tape and assumes the date it finds in that file header is the expiration date for the entire tape. Therefore, if you are writing an .oaij file to tape, specifying the Tape_Expiration qualifier only has meaning if the .oaij file is the first file on the tape. You can guarantee that the .oaij file will be the first file on the tape by specifying the Rewind qualifier and overwriting any existing files on the tape. When the first file on the tape contains an expiration date in the file header, you cannot overwrite the tape before the expiration date unless you have the OpenVMS SYSPRV or BYPASS privilege. Similarly, when you attempt to perform a recover operation with an .oaij file on tape, you cannot perform the recover operation after the expiration date recorded in the first file on the tape unless you have the OpenVMS SYSPRV or BYPASS privilege By default, no expiration date is written to the .oaij file header. In this case, if the .oaij file is the first file on the tape, the tape can be overwritten immediately. If the .oaij file is not the first file on the tape, the ability to overwrite the tape is determined by the expiration date in the file header of the first file on the tape. You cannot explicitly set a tape expiration date for an entire volume. The volume expiration date is always determined by the expiration date of the first file on the tape. The Tape_ Expiration qualifier cannot be used with a backup operation to disk. 4 Trace Trace Notrace Specifies that the optimization of the .aij file be traced. The default is the Notrace qualifier, where optimization is not traced. When optimization is traced, the output from the Trace qualifier identifies transactions in the .aij file by transaction sequence numbers (TSNs) and describes what Oracle RMU did with each transaction during the optimization process. You can specify the Log qualifier with the Trace qualifier. 3 Usage_Notes o To use the RMU Optimize After_Journal command for a database, you must have the RMU$BACKUP or RMU$RESTORE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o You cannot optimize an .aij file in the process of backing it up. You must first back up the .aij file, using the RMU Backup After_Journal command with the Format=Old_File qualifier, and then optimize it. o As part of the optimization process, Oracle RMU sorts journal records by physical dbkey which improves I/O performance of the recovery. Because AIJ file optimization uses the OpenVMS Sort/Merge utility (SORT/MERGE) to sort journal records, you can improve the efficiency of the sort operation by changing the number and location of the work files used by SORT/MERGE. The number of work files is controlled by the RDMS$BIND_SORT_ WORKFILES logical name. The allowable values are 1 through 10 inclusive, with a default value of 2. The location of these work files can be specified with device specifications, using the SORTWORKn logical name (where n is a number from 0 to 9). See the OpenVMS documentation set for more information on using SORT/MERGE. See the Oracle Rdb7 Guide to Database Performance and Tuning for more information on using these Oracle Rdb logical names. o Do not use the OpenVMS Alpha High Performance Sort/Merge utility (selected by defining the logical name SORTSHR to SYS$SHARE:HYPERSORT) when using the RMU Optimize After_Journal command. HYPERSORT does not support several of the interfaces the command uses. In addition, HYPERSORT does not report errors or warnings when it is used with the RMU Optimize After_Journal command. Make sure that the SORTSHR logical name is not defined to reference HYPERSORT.EXE. o You can redirect the AIJ rollforward temporary work files and the database recovery (DBR) redo temporary work files to a different disk and directory location than the default (SYS$DISK) by assigning a different directory to the RDM$BIND_ AIJ_WORK_FILE logical in the LNM$FILE_DEV name table and a different directory to the RDM$BIND_DBR_WORK_FILE logical in the LNM$SYSTEM_TABLE, respectively. This can be helpful in alleviating I/O bottlenecks that might be occurring in the default location. o You can optimize an inactive .aij file that results, for example, from backing up and renaming an extensible .aij file. Backing up and renaming an extensible .aij file creates a new active, primary .aij file and makes the previous .aij file inactive. After optimizing the inactive .aij file, you can use the OpenVMS BACKUP command to back up the .oaij file. Note that you cannot optimize an active, primary .aij file. o The RMU Optimize After_Journal command can read an .aij file on disk or a backed up .aij file on disk or on tape that is in the Old_File format, and it can write the .oaij file to disk or to tape in either Old_File or New_Tape format. o If an RMU Optimize After_Journal command is issued from a batch job, tape requests and problems are reported to the tape operator. This occurs because tape requests and problems often require manual intervention, and if the RMU Optimize After_Journal command was issued from a batch job, the only available person might be the operator. o When the RMU Optimize After_Journal command is issued interactively and a tape request or problem arises, Oracle RMU notifies the person who issued the command through the I/O channel assigned to the logical name SYS$COMMAND. After being notified of the problem, the user who issued the command can either fix the problem (if the user has access to the tape drive) or contact the tape operator to ask the tape operator to fix the problem. The REQUEST command can be used to notify the tape operator, as follows: $ REQUEST/REPLY/TO=TAPES - _$ "Please Write Enable tape ATOZBG on drive $255$MUA6:" o You should use the density values added in OpenVMS Version 7.2-1 for OpenVMS tape device drivers that accept them because previously supported values may not work as expected. If previously supported values are specified for drivers that support the OpenVMS Version 7.2-1 density values, the older values are translated to the Version 7.2-1 density values if possible. If the value cannot be translated, a warning message is generated, and the specified value is used. If you use density values added in OpenVMS Version 7.2-1 for tape device drivers that do not support them, the values are translated to acceptable values if possible. If the value cannot be translated, a warning message is generated and the density value is translated to the existing default internal density value (MT$K_DEFAULT). One of the following density-related errors is generated if there is a mismatch between the specified density value and the values that the tape device driver accepts: %DBO-E-DENSITY, TAPE_DEVICE:[000000]DATABASE.BCK; does not support specified density %DBO-E-POSITERR, error positioning TAPE_DEVICE: %DBO-E-BADDENSITY, The specified tape density is invalid for this device o If you want to use an unsupported density value, use the VMS INITIALIZE and MOUNT commands to set the tape density. Do not use the Density qualifier. o Because data stream names representing the database are generated based on the backup file name specified for the Oracle RMU backup command, you must either use a different backup file name to store the next backup of the database to the Librarian utility or first delete the existing data streams generated from the backup file name before the same backup file name can be reused. To delete the existing data streams stored in the Librarian utility, you can use a Librarian management utility or the Oracle RMU Librarian/Remove command. 3 Examples Example 1 The following command creates an .oaij file named mf_ personnel.oaij from the .aij file named mf_personnel.aij: $ RMU/OPTIMIZE/AFTER_JOURNAL MF_PERSONNEL.AIJ MF_PERSONNEL.OAIJ Example 2 The following example uses a density value with compression: RMU/OPTIMIZE/AFTER_JOURNAL /DENSITY=(TK89,COMPACTION)/REWIND - /LABEL=(LABEL1,LABEL2) MF_PERSONNEL.AIJ TAPE1:MF_PERSONNEL.OAIJ, TAPE2: 2 Populate_Cache Reads one or more tables and indexes from the database and stores the data rows or index nodes in caches if they exist. 3 Description The RMU Populate_Cache command allows one or more tables and indexes to be read from the database and stored in caches if they exist. Sorted indexes are read top-down, one index level at a time. Hashed indexes are read by sequentially scanning the storage areas containing the hashed indexes and fetching all nodes and the system record from each database page. Data table rows are read by sequentially scanning the storage areas containing the table and fetching all rows of the relation. 3 Format (B)0RMU/Populate_Cache root-file-spec Command Qualifiers x Defaults x /Index=index-list x None /[No]Log x Current DCL verify switch /[No]Only_Cached x /Only_Cached /Statistics_Interval=n x /Statistics_Interval=10 /Table=table-list x None /Transaction_Type=transaction-mode x /Transaction-Type=Automatic 3 Parameters 4 root-file-spec Specifies the database root (.rdb) file. The default file type is .rdb. 3 Command_Qualifiers 4 Index Index=Index-list Specifies one or more indexes to fetch. All nodes are fetched from each index. If you list multiple indexes, separate the index names with a comma and enclose the list within parentheses. Wildcard characters asterisk (*) and percent sign (%) are allowed. 4 Log Log Nolog Specifies whether the processing of the command is reported to SYS$OUTPUT. Specify the Log qualifier to request that information about the operation be displayed, and the Nolog qualifier to prevent it. If you specify neither qualifer, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 4 Only_Cached Only_Cached Noonly_Cached Specifies whether table or index content is to be read only if the table or index has an associated row cache. The default is to read data only from objects that have a cache. If the Noonly_ Cached qualifier is specified, then all data from the specified tables or indexes is read. 4 Statistics_Interval Statistics_Interval=n Specifies if statistics information is to be periodically displayed during the populate operation. The default for this qualifier is an interval of 10 seconds. If you do not use this qualifier no statistics are displayed. 4 Table Table=table-list Specifies one or more tables to be processed. All rows are fetched from each table. If you list multiple tables, separate the table names with a comma, and enclose the list within parentheses. Wildcard characters asterisk (*) and percent sign (%) are allowed. 4 Transaction_Type Transaction_Type=option Allows you to specify the transaction mode for the transactions used to perform the analyze operation. Valid options are: o Automatic o Read_Only o Noread_Only You must specify an option if you use this qualifier. If you do not specify any form of this qualifier, the Transaction_Type=Automatic qualifier is the default. This qualifier specifies that Oracle RMU is to determine the transaction mode. The Transaction_Type=Read_Only qualifier specifies the transactions used to perform the analyze operation be set to read-only mode. When you explicitly set the transaction type to read-only, snapshots need not be enabled for all storage areas in the database, but must be enabled for those storage areas that are read. Otherwise, you receive an error and the analyze operation fails. You might select this option if not all storage areas have snapshots enabled and you are analyzing objects that are stored only in storage areas with snapshots enabled. In this case, using the Transaction_Type=Read_Only qualifier allows you to perform the analyze operation and impose minimal locking on other users of the database. The Transaction_Type=Noread_Only qualifier specifies that the transactions used to for the analyze operation be set to read/write mode. You might select this option if you want to eradicate the growth of snapshot files that occurs during a read- only transaction and are willing to incur the cost of increased locking that occurs during a read/write transaction. 2 Reclaim Allows you to rapidly reclaim deleted dbkeys and locked space from database pages. 3 Description Applications that specify the database attach attribute DBKEY SCOPE IS ATTACH can accumulate locked space and locked dbkeys within the database. If one user is connected to the database in DBKEY SCOPE IS ATTACH mode, all users are forced to operate in this mode, even if they are explicitly connected in TRANSACTION mode. No dbkeys are reused until the ATTACH session disconnects. The RMU Reclaim command allows database keys of deleted rows to be rapidly reset in one or more storage areas. The RMU Reclaim command reads and updates all pages in a storage area, and, where possible, releases locked lines and locked free space so that they are available for later allocation. 3 Format (B)0RMU/Reclaim root-file-spec Command Qualifiers x Defaults x /Area[=storage-area-list] x All storage areas /[No]Log x /NoLog 3 Parameters 4 root-file-spec Specifies the database that contains locked areas or keys to be reclaimed. The default file extension is .rdb. 3 Command_Qualifiers 4 Area Area=storage-area-list Lists the storage areas to be reclaimed. The default is all storage areas. 4 Log Log NoLog Displays a log message as each storage area is reclaimed. The default is Nolog. 3 Usage_Notes o The RMU Reclaim command runs on-line and does not require exclusive access. However, if there are any users connected to the database in DBKEY SCOPE IS ATTACH mode, the RMU/RECLAIM operation has greatly reduced effect. In order to release all possible locked space, there should be no users attached to the database in DBKEY SCOPE IS ATTACH mode. o To allow database page locked space to be reclaimed, the database session that controls the locked space must be detached from the database. This can be accomplished by having each attached session disconnect and reconnect to the database. 2 Recover Completes a database reconstruction by processing past transactions from the after-image journal (.aij) file or optimized after-image journal (.oaij) file against a database restored from a backup file. 3 Description You can use the RMU Recover command to apply the contents of an .aij file to a restored copy of your database. Oracle RMU rolls forward the transactions in the .aij file into the restored copy of the database. The RMU Recover command accepts a list of .aij or .oaij file names. Unless you specify the Noautomatic qualifier, the RMU Recover command attempts to automatically complete the recovery operation by applying the journals currently associated with the database in the current journal configuration if they are in the recovery sequence. For example, if you specify the following RMU Recover command, Oracle RMU not only recovers AIJ1, but also AIJ2, AIJ3, and so on, for all journals in the recovery sequence: $ RMU/RECOVER AIJ1 However, note that this automatic recovery feature means that if you want to specify a termination condition, you must specify the Until qualifier. Example 1 demonstrates how to specify a termination condition with the Until qualifier. If you are using extensible journals, you can also use the RMU Backup After_Journal command to copy your database's .aij file to tape, and truncate the original .aij file without shutting down your database. If you have backed up your .aij files (using the RMU Backup After_Journal command), these .aij files are no longer part of the current journal configuration and automatic recovery does not take place because Oracle RMU does not know where to find the .aij files. (There is one exception to this rule: if the only .aij file that has been backed up is the first .aij file in the recovery sequence, then automatic recovery occurs. You specify the backed up .aij file on the Oracle RMU command line and Oracle RMU can determine where the remaining on-disk .aij files reside.) When automatic recover does not, or cannot occur, you must specify the complete list of .aij files on the RMU Recover command line to return your database to the desired state. If your backup files were created using the Noquiet_Point qualifier, you must provide the names of all the .aij files in just one command. In addition, you must be careful to apply these .aij files to the database in the order in which they were created. Oracle RMU checks the validity of the journal file entries against your database and applies only appropriate transactions. If none of the transactions apply, you will receive a warning message. You can access your database for retrieval of data between recovery steps, but you must not perform additional updates if you want to perform more recovery steps. If a system failure causes a recovery step to abort, you can simply issue the RMU Recover command again. Oracle RMU scans the .aij file until it finds the first transaction that has not yet been applied to your restored database. Oracle RMU begins recovery at that point. 3 Format (B)0RMU/Recover aij-file-name-list Command Qualifiers x Defaults x /Active_IO=max-reads x /Active_IO=3 /Aij_Buffers=integer x /Aij_Buffers=20 /Areas [= storage-area[,...] ] x All storage areas /[No]Automatic x /Automatic /[No]Confirm[=options] x See description /Encrypt=({Value=|Name=}[,Algorithm=]) x See description /Format={Old_File|New_Tape} x /Format=Old_File /Just_Corrupt x See description /Label=(label-name-list) x See description /Librarian[=options] x None /[No]Log x Current DCL verify value /[No]Media_Loader x See description /[No]Online x /Noonline /Order_Aij_Files x See description /Output=file-name x See description /Prompt={Automatic|Operator|Client} x See description (B)0/Resolve x See description /[No]Rewind x /Norewind /Root=root-file-name x See description /[No]Trace x See Description /Until=date-time x Current time 3 Parameters 4 aij-file-name-list A list of after-image journal (.aij) files to be applied to the database. You can supply this list using one of the following methods: o List the .aij files on the command line in the order in which they were created. In other words, the oldest .aij file must be the first in the list. o Use an asterisk (*) or percent sign (%) to represent the .aij files. The .aij files are processed in the order that they are presented by OpenVMS. o Append all your .aij files into one file and supply a single .aij file name. You must be certain when you append the files that you append them in the order in which they were created. o Use an indirect command file. Include an .aij file name on each line of the command file. If the number of .aij files needed for recovery is large, listing each one on the command line can exceed the maximum allowed command-line length. You can avoid this problem by using an indirect command file. See the Indirect-Command-Files help entry for more information on indirect command files. 3 Command_Qualifiers 4 Active_IO Active_IO=max-reads Specifies the maximum number of read operations from a backup device that the RMU Recover command attempts simultaneously. This is not the maximum number of read operations in progress; that value is a function of active system I/O operations. The value of the Active_IO qualifier can range from 1 to 5. The default value is 3. Values larger than 3 can improve performance with some tape drives. 4 Aij_Buffers Aij_Buffers=integer Specifies the number of buffers to be used by the recovery process. The default is 20 buffers. The valid range is 2 to 1,048,576 buffers. If the database root file is available, you can use the RMU Dump After_ Journal command with the Option=Statistics qualifier to find a recommended value for this qualifier. See Dump After_journal for details. 4 Areas Areas[=storage-area[,...]] Specifies the areas you want to recover. You can specify each storage area by name or by the area's ID number. You should use the Areas qualifier only if you have inconsistent storage areas to recover. The default for the Areas qualifier is all storage areas that must be recovered to make the database consistent. If the Areas qualifier is specified, a recovery operation by area recovers until the storage areas being rolled forward are current with the other storage areas, then recovery stops, regardless of the time specified by the Until qualifier. When the Areas qualifier is not specified or the Areas=* qualifier is specified, Oracle RMU recovers all the storage areas in the database to the time specified by the Until qualifier or to the time of the last committed transaction in the .aij file that can be applied. When the Areas qualifier is specified without a value, Oracle RMU recovers to the earliest consistent state only those storage areas that are not current with the database root (.rdb) file of the database. The Areas qualifier works in the following manner: o If the Areas qualifier is specified without a value, Oracle RMU automatically determines what areas are inconsistent and recovers those areas. If an inconsistent area cannot be recovered because it is at a higher transaction sequence number (TSN) value than the database root file, the entire database is recovered even if the Areas qualifier was specified. See the Oracle Rdb Guide to Database Maintenance for information on TSNs. o If the Areas qualifier is omitted or the Areas qualifier is specified as Areas=*, the entire database (all storage areas) is recovered. o If the Areas qualifier is specified as Areas=(A1,A2,A3), only areas A1, A2, and A3 are recovered until they are consistent. If one of these areas is already consistent, or if an area is at a higher TSN value than the database root file, the entire database is recovered. o If the Online qualifier is specified with the Areas qualifier (as in the first three list items) and the end result is that the entire database must be recovered, an error message is generated because you can recover only individual areas by using the Online qualifier, not the entire database. You cannot use the Areas qualifier with the Just_Corrupt qualifier because the Areas qualifier implies recovery for all named areas and pages in those areas. (That is, use of the Just_ Corrupt qualifier with the Areas qualifier is redundant.) The Areas qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 4 Automatic Automatic Noautomatic Specifies whether or not Oracle RMU should attempt automatic recovery of .aij files. If you specify the Noautomatic qualifier, only the .aij file or files you list on the Oracle RMU command line are applied. If you specify the Automatic qualifier, Oracle RMU attempts to recover all the .aij files currently associated with the database. The Automatic qualifier is the default; Oracle RMU attempts to recover all the .aij files currently associated with the database unless the .aij files have been backed up. See the description section for more information on how automatic recovery works. 4 Confirm Confirm[=options] Noconfirm Specifies whether or not the RMU /RECOVER command causes the operator to be queried when an incorrect sequence of AIJ files is detected. The default for interactive recoveries is /CONFIRM, which prompts the user to see if he wants to continue. The default for RMU/RECOVER/NOCONFIRM and RMU/RECOVER executed in batch jobs is to terminate the RMU/RECOVER at the point where the out of sequence AIJ file is detected (equivalent to RMU/RECOVER/CONFIRM=ABORT). To override the default behavior, the user can continue to roll forward and ignore the missing AIJ file either by specifying the command syntax RMU/RECOVER/CONFIRM to get a prompt on whether to continue rolling forward if there is an AIJ sequence gap, or by specifying the syntax RMU/CONFIRM=CONTINUE if he does not want the prompt or is executing the RMU/RECOVER in a batch job. NOTE Oracle recommends that, in general, an incorrect journal sequence not be applied as a corrupt database may result. The /Order_Aij_Files qualifier can be used to help ensure that the specified journals are applied in the correct order. The Confirm qualifier accepts the following options: o CONFIRM=CONTINUE Do not prompt the user if a sequence gap is detected on the next AIJ file to be rolled forward but ignore the missing AIJ file and continue rolling forward. o CONFIRM=ABORT Do not prompt the user if a sequence gap is detected on the next AIJ roll forward but end the database recover at this point. This is the same as the default behavior for RMU/RECOVER/NOCONFIRM and RMU/RECOVER in batch. 4 Encrypt Encrypt=({Value=|Name=}[,Algorithm=]) The Encrypt qualifier is used to recover the database from an encrypted after image journal backup file. Specify a key value as a string or, the name of a predefined key. If no algorithm name is specified the default is DESCBC. For details on the Value, Name and Algorithm parameters see HELP ENCRYPT. This feature requires the OpenVMS Encrypt product to be installed and licensed on this system. This feature only works for a newer format backup file which has been created using the Format=New_Tape qualifier. Therefore you have to specify the Format=New_Tape qualifier with this command if you also use the Encrypt qualifier. Synonymous with the Format=Old_File and Format=New_Tape qualifiers. See the description of those qualifiers. 4 Format Format=Old_File Format=New_Tape Specifies whether the backed up or optimized .aij file was written in the old (disk-optimized) or the new (tape-optimized) format. The Format=Old_File qualifier is the default. You must specify the same Format qualifier that was used with the RMU Backup After_Journal command or the RMU Optimize After_Journal command. If your .aij file resides on disk, you should use the Format=Old_File qualifier. If you specified the Format=Old_File qualifier when you optimized or backed up the .aij file to tape, you must mount the backup media by using the DCL MOUNT command before you issue the RMU Recover command. Because the RMU Recover command will use RMS to read the tape, the tape must be mounted as an OpenVMS volume (that is, do not specify the /FOREIGN qualifier with the MOUNT command). If you specify the Format=New_Tape qualifier, you must mount the backup media by using the DCL MOUNT /FOREIGN command before you issue the RMU Recover command. Similarly, if you specify OpenVMS access (you do not specify the /FOREIGN qualifier on the DCL MOUNT command) although your .aij backup was created using the Format=New_Tape qualifier, you will receive an RMU-F-MOUNTFOR error. The following tape qualifiers have meaning only when used in conjunction with the Format=New_Tape qualifier: Active_IO Label Rewind 4 Just_Corrupt Just_Corrupt Specifies that only inconsistent pages in the corrupt page table (CPT) and areas marked as inconsistent should be recovered. You can use this qualifier while users are attached to the database. You can use the Just_Corrupt qualifier with the Until qualifier to limit the recovery period to a particular point in time. You cannot use the Areas qualifier with the Just_Corrupt qualifier because the Areas qualifier implies recovery for all named areas and pages in those areas. (That is, use of the Just_ Corrupt qualifier with the Areas qualifier is redundant.) If you do not specify the Just_Corrupt qualifier, all pages are recovered. 4 Just_Pages Just_Pages This qualifier is replaced with the Just_Corrupt qualifier beginning in Oracle Rdb V7.0. See the description of the Just_ Corrupt qualifier. Specifies the 1- to 6-character string with which the volumes of the backup 4 Label Label=(label-name-list) Specifies the 1- to 6-character string with which the volumes of the backup file have been labeled. The Label qualifier is applicable only to tape volumes. You must specify one or more label names when you use the Label qualifier. You can specify a list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas, and enclose the list of names within parentheses. In a normal recovery operation, the Label qualifier you specify with the RMU Recover command should be the same Label qualifier you specified with the RMU Backup After_Journal command to back up your .aij files. The Label qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 4 Librarian Librarian=options Use the Librarian qualifier to restore files from data archiving software applications that support the Oracle Media Management interface. The file name specified on the command line identifies the stream of data to be retrieved from the Librarian utility. If you supply a device specification or a version number it will be ignored. Oracle RMU supports retrieval using the Librarian qualifier only for data that has been previously stored by Oracle RMU using the Librarian qualifer. The Librarian qualifier accepts the following options: o Trace_file=file-specification The Librarian utility writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian utility. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian utility. The pre-determined values are : - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical_name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian utility can use to specify catalogs or archives where Oracle Rdb backup files are stored, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of any Oracle RMU command that accesses the Librarian application. The following OpenVMS logical names must be defined for use with a Librarian utility before you execute an Oracle RMU backup or restore operation. Do not use the Logical_Names option provided with the Librarian qualifier to define these logical names. o RMU$LIBRARIAN_PATH This logical name must be defined so that the shareable Librarian image can be loaded and called by Oracle RMU backup and restore operations. The translation must include the file type (for example, .exe), and must not include a version number. The shareable Librarian image must be an installed (known) image. See the Librarian utility documentation for the name and location of this image and how it should be installed. o RMU$DEBUG_SBT This logical name is not required. If it is defined, Oracle RMU will display debug tracing information messages from modules that make calls to the Librarian shareable image. You cannot use device specific qualifiers such as Rewind, Density, or Label with the Librarian qualifier because the Librarian utility handles the storage meda, not Oracle RMU. 4 Log Log Nolog Specifies that the recovery activity be logged. The default is the setting of the DCL VERIFY flag, which is controlled by the DCL SET VERIFY command. When recovery activity is logged, the output from the Log qualifier provides the number of transactions committed, rolled back, and ignored during the recovery process. You can specify the Trace qualifier with the Log qualifier. 4 Media_Loader Media_Loader Nomedia_Loader Use the Media_Loader qualifier to specify that the tape device from which the .aij file is being read has a loader or stacker. Use the Nomedia_Loader qualifier to specify that the tape device does not have a loader or stacker. By default, if a tape device has a loader or stacker, Oracle RMU should recognize this fact. However, occasionally Oracle RMU does not recognize that a tape device has a loader or stacker. Therefore, when the first tape has been read, Oracle RMU issues a request to the operator for the next tape, instead of requesting the next tape from the loader or stacker. Similarly, sometimes Oracle RMU behaves as though a tape device has a loader or stacker when actually it does not. If you find that Oracle RMU is not recognizing that your tape device has a loader or stacker, specify the Media_Loader qualifier. If you find that Oracle RMU expects a loader or stacker when it should not, specify the Nomedia_Loader qualifier. 4 Online Online Noonline Specifies that the recover operation be performed while other users are attached to the database. The Online qualifier can only be used with the Area or Just_Corrupt qualifier. The areas or pages to be recovered are locked for exclusive access, so the operation is not compatible with other uses of the data in the areas or on the pages specified. The default is the Noonline qualifier. 4 Order_Aij_Files Specifies that the input after-image journal files are to be processed in ascending order by sequence number. The .aij files are each opened, the first block is read to determine the sequence number, and the files are closed prior to the sequence number sorting operation. The Order_Aij_Files can be especially useful if you use wildcards to specify .aij files. The Order_Aij_Files qualifier can also eliminate some .aij files from processing if they are known to be prior to the database recovery sequence starting point. Note that due to the fact that the .aij backup files might have more than one journal sequence in them, it is not always possible for RMU to eliminate every journal file that might otherwise appear to be unneeded. But for those journals where RMU is able to know for certain that the journal will not be needed based on the database recovery restart information, journals can be avoided from having to be processed. 4 Output Output=file-name Redirects the log and trace output (selected with the Log and Trace qualifiers) to the named file. If this qualifier is not specified, the output generated by the Log and Trace qualifiers, which can be voluminous, is displayed on your terminal. 4 Prompt Prompt=Automatic Prompt=Operator Prompt=Client Specifies where server prompts are to be sent. When you specify Prompt=Automatic, prompts are sent to the standard input device, and when you specify Prompt=Operator, prompts are sent to the server console. When you specify Prompt=Client, prompts are sent to the client system. 4 Resolve Resolve Recovers a corrupted database and resolves an unresolved transaction by completing the transaction. See the help entry for the RMU Recover Resolve command for a description of the options available with the Resolve qualifier. 4 Rewind Rewind Norewind Specifies that the tape that contains the backup file be rewound before processing begins. The tape is searched for the backup file starting at the beginning-of-tape (BOT). The Norewind qualifier is the default and causes the backup file to be searched starting at the current tape position. The Rewind and Norewind qualifiers are applicable only to tape devices. Oracle RMU returns an error message if these qualifiers are used and the target device is not a tape device. 4 Root Root=root-file-name Specifies the name of the database to which the journal should be applied. The Root qualifier allows you to specify a copy of a database instead of the original whose file specification is in the .aij file. Use the Root qualifier to specify the new location of your restored database root (.rdb) file. Specifying this qualifier lets you roll forward a database copy (possibly residing on a different disk) by following these steps: 1. Use the RMU Backup command to make a backup copy of the database: $ RMU/BACKUP MF_PERSONNEL.RDB MF_PERS_FULL_BU.RBF This command writes a backup file of the database mf_personnel to the file mf_pers_full_bu.rbf. 2. Use the RMU Restore command with the Root and Directory qualifiers, stating the file specifications of the database root and storage area files in the database copy. $ RMU/RESTORE/ROOT=DB3:[USER]MF_PERSONNEL/DIRECTORY=DB3:[USER] - _$ MF_PERS_FULL_BU This command restores the database on disk DB3: in the directory [USER]. Default file names and file extensions are used. 3. If the database uses after-image journaling, you can use the RMU Recover command to roll forward the copy. $ RMU/RECOVER DBJNL.AIJ/ROOT=DB3:[USER]MF_PERSONNEL.RDB Thus, transactions processed and journaled since the backup operation are recovered on the copy on the DB3: disk. Correct operation of this procedure requires that there are no write transactions for the restored copy between the restore and recover steps. If you do not specify the Root qualifier, Oracle RMU examines the .aij file to determine the exact name of the database root (.rdb) file to which the journaled transactions will be applied. This name, which was stored in the .aij file, is the full file specification that your .rdb file had when after-image journaling was enabled. The journal file for a single-file database does not include the file name for the database; to recover a single-file database, you must specify the location of the database to be recovered by using the Root qualifier. 4 Trace Trace Notrace Specifies that the recovery activity be logged. The default is the setting of the DCL VERIFY flag, which is controlled by the DCL SET VERIFY command. When recovery activity is logged, the output from the Trace qualifier identifies transactions in the .aij file by TSN and describes what Oracle RMU did with each transaction during the recovery process. You can specify the Log qualifier with the Trace qualifier. 4 Until Until=date-time Use the Until qualifier to limit the recovery to those transactions in the journal file bearing a starting timestamp no later than the specified time. For example, suppose your database fails today, but you have reason to believe that something started to go wrong at noon yesterday. You might decide that you only want to restore the database to the state it was in as of noon yesterday. You could use the Until qualifier to specify that you only want to recover those transactions that have a timestamp of noon yesterday or earlier. If you do not specify the Until qualifier, all committed transactions in the .aij file will be applied to your database. If you specify the Until qualifier, but do not specify a date- time, the current time is the default. If the Until qualifier is specified with a recover-by-area operation, the operation terminates when either the specified time is reached in the transaction sequence or the specified storage areas become consistent with the other storage areas; whichever condition occurs first. 3 Usage_Notes o To use the RMU Recover command for a database, you must have the RMU$RESTORE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o You can use the RMU Backup After_Journal command to copy an extensible .aij file to tape and truncate the original .aij file without shutting down your database. o Transactions are applied to the restored copy of your database in the order indicated by their commit sequence number and the commit record in the .aij file; timestamps are not used for this purpose. Therefore, you need not be concerned over time changes made to the system (for example, resetting the time for United States daylight saving time) or inconsistencies in the system time on different nodes in a cluster. The only occasion when timestamps are considered in the application of .aij files is when you specify the Until qualifier. In this case, the timestamp is used only as the point at which to stop the recovery, not as a means to serialize the order in which transactions are applied. See the description of the Until qualifier for more information. o You can redirect the AIJ rollforward temporary work files and the database recovery (DBR) redo temporary work files to a different disk and directory location than the default (SYS$DISK) by assigning a different directory to the RDM$BIND_ AIJ_WORK_FILE logical in the LNM$FILE_DEV name table and a different directory to the RDM$BIND_DBR_WORK_FILE logical in the LNM$SYSTEM_TABLE, respectively. This can be helpful in alleviating I/O bottlenecks that might be occurring in the default location. o In a normal recovery operation, the Format and Label qualifiers you specify with the RMU Recover command should be the same Format and Label qualifiers you specified with the RMU Backup After_Journal command to back up your .aij files or with the RMU Optimize After_Journal command to optimize your .aij files. For more information on the type of access to specify when mounting tapes, see the description of the Format=Old_File and Format=New_Tape qualifiers in the Format section. o The following restrictions apply to using optimized .aij files with recovery operations: - Optimized .aij files cannot be used as part of by-area recovery operations (recovery operations that use the RMU Recover command with the Area qualifier). - Optimized .aij files cannot be used as part of by-page recovery operations (recovery operations that use the RMU Recover command with the Just_Corrupt qualifier). - Optimized .aij files cannot be specified for an RMU Recover command that includes the Until qualifier. The optimized .aij file does not retain enough of the information from the original .aij file for such an operation. - Optimized .aij files cannot be used with a recovery operation if the database has been modified since the .aij file was optimized. The workaround for these restrictions against using optimized .aij files in recovery operations is to use the original, unoptimized .aij file in the recovery operation instead. o You can read your database between recovery steps, but you must not perform additional updates if you want to perform more recovery steps. o If a system failure causes a recovery step to abort, you can simply issue the RMU Recover command again. Oracle RMU scans the .aij file until it finds the first transaction that has not yet been applied to your restored database. Oracle RMU begins recovery at that point. o You can use the RMU Recover command to apply the contents of an .aij file to a restored copy of your database. Oracle RMU will roll forward the transactions in the .aij file into the restored copy of the database. You can use this feature to maintain an up-to-date copy of your database for fast recovery after a failure. To do this, use the RMU Recover command to periodically apply your .aij files to a separate copy of the database. When you employ this procedure for fast recovery, you must be absolutely certain that no one will execute an update transaction on the database copy. Should someone execute an update transaction, it might result in the inability to apply the .aij files correctly. o See the Oracle Rdb Guide to Database Maintenance for information on the steps Oracle RMU follows in tape label checking. o When you use an optimized after-image journal for recovery, the optimal number of buffers specified with the Aij_Buffers qualifier depends on the number of active storage areas being recovered. For those journals optimized with Recover_ Method=Sequential, a buffer count of 250 to 500 is usually sufficient. When you use journals optimized with Recover_Method=Scatter, reasonable performance can usually be attained with a buffer count of about five times the number of active storage areas being recovered (with a minimum of about 250 to 500 buffers). o The number of asynchronous prefetch (APF) buffers is also a performance factor during recovery. For recovery operations of optimized after-image journals, the RMU Recover command sets the number of APF buffers (also known as the APF depth) based on the values of the process quotas ASTLM, BYTLM, and the specified AIJ_Buffers value. The APF depth is set to the maximum of: - 50% of the ASTLM process quota - 50% of the DIOLM process quota - 25% of the specified AIJ_Buffers value The accounts and processes that perform RMU Recover operations should be reviewed to ensure that various quotas are set to ensure high levels of I/O performance. The following table lists suggested quota values for recovery performance. Quota Setting DIOLM Equal to or greater than half of the count of database buffers specified by the AIJ_Buffers qualifier. Miminum of 250. BIOLM Equal to or greater than the setting of DIOLM. ASTLM Equal to or greater than 50 more than the setting of DIOLM. BYTLM Equal to or greater than 512 times the database buffer size times one half the value of database buffers specified by the AIJ_Buffers qualifier. Based on a 12-block buffer size and the desire to have 100 asynchronous I/O requests outstanding (either reading or writing), the minimum suggested value is 614,400 for a buffer count of 200. WSQUOTA Large enough to avoid excessive page faulting. WSEXTENT FILLM 50 more than the count of database storage areas and snapshot storage areas. 3 Examples Example 1 In the following example, the RMU Recover command requests recovery from the .aij file personnel.aij located on PR$DISK in the SMITH directory. It specifies that recovery should continue until 1:30 P.M. on May 7, 1996. Because the Trace qualifier is specified, the RMU Recover command displays detailed information about the recovery operation to SYS$OUTPUT. $ RMU/RECOVER/UNTIL="07-MAY-1996 13:30"/TRACE PR$DISK:[SMITH]PERSONNEL %RMU-I-LOGRECDB, recovering database file DISK1:[DB.70]MF_PERSONNEL.RDB;1 %RMU-I-LOGRECSTAT, transaction with TSN 0:256 committed %RMU-I-AIJONEDONE, AIJ file sequence 0 roll-forward operations completed %RMU-I-AIJAUTOREC, starting automatic after-image journal recovery %RMU-I-AIJONEDONE, AIJ file sequence 1 roll-forward operations completed %RMU-W-NOTRANAPP, no transactions in this journal were applied %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJFNLSEQ, to start another AIJ file recovery, the sequence number needed will be 1 Example 2 The following example shows how to use .aij files to recover a database: SQL> CREATE DATABASE FILENAME DISK1:[SAMPLE]TEST_DB cont> RESERVE 5 JOURNALS; SQL> -- SQL> -- Use the DISCONNECT ALL statement to detach from the database, SQL> -- then issue the ALTER DATABASE statement that automatically SQL> -- invokes the specified database. SQL> -- SQL> DISCONNECT ALL; SQL> -- SQL> -- Create after-image journaling. The .aij files are given the SQL> -- names aij_one.aij and aij_two.aij (and are placed on a disk SQL> -- other than the disk holding the .rdb and .snp files): SQL> -- SQL> ALTER DATABASE FILENAME DISK1:[SAMPLE]TEST_DB cont> JOURNAL IS ENABLED cont> ADD JOURNAL AIJ_ONE cont> FILENAME 'USER$DISK:[CORP]AIJ_ONE' cont> BACKUP FILENAME 'USER$DISK2:[CORP]AIJ_ONE' cont> ADD JOURNAL AIJ_TWO cont> FILENAME 'USER$DISK3:[CORP]AIJ_TWO' cont> BACKUP FILENAME 'USER$DISK4:[CORP]AIJ_TWO'; SQL> EXIT $ ! $ ! Using the RMU Backup command, make a backup copy of the database. $ ! This command ensures that you have a copy of the $ ! database at a known time, in a known state. $ ! $ RMU/BACKUP DISK1:[SAMPLE]TEST_DB USER2:[BACKUPS]TEST_BACKUP.RBF $ ! $ ! Now you can use SQL with after-image journaling enabled. $ ! $ SQL SQL> -- SQL> -- Attach to the database and perform some data definition SQL> -- and storage. SQL> -- SQL> ATTACH 'FILENAME DISK1:[SAMPLE]TEST_DB'; SQL> CREATE TABLE TABLE1 (NEW_COLUMN CHAR(10)); SQL> INSERT INTO TABLE1 (NEW_COLUMN) VALUES ('data'); SQL> COMMIT; SQL> EXIT $ ! $ ! Imagine that a disk failure occurred here. In such a situation, $ ! the current database is inaccessible. You need a prior copy $ ! of the database to roll forward all the transactions in the $ ! .aij file. $ ! $ ! $ ! You know that the backup file of the database is $ ! uncorrupted. Use the RMU Restore command to restore and recover $ ! the database. You do not have to issue the RMU Recover command $ ! because the RMU Restore command will automatically recover the $ ! database. $ ! $ RMU/RESTORE/NOCDD_INTEGRATE/DIR=DDV21:[TEST] - _$ USER2:[BACKUPS]TEST_BACKUP.RBF %RMU-I-AIJRSTAVL, 2 after-image journals available for use %RMU-I-AIJRSTMOD, 1 after-image journal marked as "modified" %RMU-I-AIJISON, after-image journaling has been enabled %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery %RMU-I-LOGRECDB, recovering database file DDV21:[TEST]TEST_DB.RDB;1 %RMU-I-AIJAUTOREC, starting automatic after-image journal recovery %RMU-I-AIJONEDONE, AIJ file sequence 0 roll-forward operations completed %RMU-I-AIJONEDONE, AIJ file sequence 1 roll-forward operations completed %RMU-W-NOTRANAPP, no transactions in this journal were applied %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJFNLSEQ, to start another AIJ file recovery, the sequence number needed will be 1 Example 3 The following example demonstrates how the recovery operation works when there are .aij backup files to be applied. First you must restore the database by using the RMU Restore command with the Norecovery qualifier, then apply the backed up .aij file by using the RMU Recover command. Oracle RMU will complete the recovery with the .aij files that were current when the restore operation was invoked. This example assumes that three .aij files have been added to the mf_personnel database prior to the first shown backup operation and that journaling is enabled. $ ! Create a backup file of the complete and full database. $ ! $ RMU/BACKUP MF_PERSONNEL DISK1:[BACKUPS]MF_PERSONNEL_BCK.RBF $ ! $ ! Updates are made to the SALARY_HISTORY and DEPARTMENTS tables. $ ! $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL'; SQL> UPDATE SALARY_HISTORY cont> SET SALARY_END='20-JUL-1993 00:00:00.00' cont> WHERE SALARY_START='14-JAN-1983 00:00:00' cont> AND EMPLOYEE_ID='00164'; SQL> INSERT INTO DEPARTMENTS cont> (DEPARTMENT_CODE, DEPARTMENT_NAME, cont> MANAGER_ID,BUDGET_PROJECTED, BUDGET_ACTUAL) cont> VALUES ('WLNS', 'WELLNESS CENTER', '00188',0,0); SQL> COMMIT; SQL> DISCONNECT DEFAULT; SQL> EXIT $ ! $ ! Create a backup file of the .aij files. $ ! $ RMU/BACKUP/AFTER_JOURNAL MF_PERSONNEL DISK2:[BACKUP]MF_PERS_AIJBCK $ ! $ ! An additional update is made to the DEPARTMENTS table. $ ! $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL'; SQL> INSERT INTO DEPARTMENTS cont> (DEPARTMENT_CODE, DEPARTMENT_NAME, MANAGER_ID,BUDGET_PROJECTED, cont> BUDGET_ACTUAL) cont> VALUES ('facl', 'FACILITIES', '00190',0,0); SQL> COMMIT; SQL> DISCONNECT DEFAULT; SQL> EXIT; $ $ ! Assume the disk holding the SALARY_HISTORY and DEPARTMENTS $ ! storage areas is lost. Restore only those areas. Specify $ ! the Norecovery qualifier since you will need to apply the $ ! .aij backup file. $ $ RMU/RESTORE/AREA DISK1:[BACKUPS]MF_PERSONNEL_BCK.RBF - _$ SALARY_HISTORY, DEPARTMENTS/NORECOVER $ ! $ ! Now recover the database. Although you only specify the .aij $ ! backup file, Oracle RMU will automatically continue the $ ! recovery with the current journals in the recovery sequence after $ ! the backed up .aij files have been applied. $ ! $ RMU/RECOVER/LOG DISK2:[BACKUP]MF_PERS_AIJBCK %RMU-I-AIJBADAREA, inconsistent storage area DISK3:[STO_AREA] DEPARTMENTS.RDA;1 needs AIJ sequence number 0 %RMU-I-AIJBADAREA, inconsistent storage area DISK3:[STO_AREA]SALARY_HISTORY.RDA;1 needs AIJ sequence number 0 %RMU-I-LOGRECDB, recovering database file DISK3:[DATABASE]MF_PERSONNEL.RDB;1 %RMU-I-LOGOPNAIJ, opened journal file DISK2:[BACKUP]MF_PERS_AIJBCK.AIJ;1 %RMU-I-AIJONEDONE, AIJ file sequence 0 roll-forward operations completed %RMU-I-LOGRECOVR, 3 transactions committed %RMU-I-LOGRECOVR, 0 transactions rolled back %RMU-I-LOGRECOVR, 0 transactions ignored %RMU-I-AIJNOACTIVE, there are no active transactions %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJNXTSEQ, to continue this AIJ file recovery, the sequence number needed will be 1 %RMU-I-AIJAUTOREC, starting automatic after-image journal recovery %RMU-I-LOGOPNAIJ, opened journal file DISK4:[CORP]AIJ_TWO.AIJ;1 %RMU-I-AIJONEDONE, AIJ file sequence 1 roll-forward operations completed %RMU-I-LOGRECOVR, 2 transactions committed %RMU-I-LOGRECOVR, 0 transactions rolled back %RMU-I-LOGRECOVR, 0 transactions ignored %RMU-I-AIJNOACTIVE, there are no active transactions %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJNXTSEQ, to continue this AIJ file recovery, the sequence number needed will be 2 %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-LOGSUMMARY, total 5 transactions committed %RMU-I-LOGSUMMARY, total 0 transactions rolled back %RMU-I-LOGSUMMARY, total 0 transactions ignored %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJGOODAREA, storage area DISK3:[STO_AREA]DEPARTMENTS.RDA;1 is now consistent %RMU-I-AIJGOODAREA, storage area DISK3:[STO_AREA]SALARY_HISTORY.RDA;1 is now consistent %RMU-I-AIJFNLSEQ, to start another AIJ file recovery, the sequence number needed will be 2 $ ! $ ! Database is restored and recovered and ready to use. $ ! Example 4 The following example demonstrates how to recover all the known inconsistent pages in a database. Assume the RMU Show Corrupt_ Pages command reveals that page 60 in the EMPIDS_LOW storage area is inconsistent and pages 11 and 123 in the EMPIDS_MID storage area is inconsistent. The RMU Recover command is issued to recover on line all pages logged inconsistent in the corrupt page table (CPT). After the recovery operation, the CPT will be empty. $ RMU/RECOVER/JUST_CORRUPT/ONLINE/LOG MF_PERSONNEL.AIJ %RMU-I-AIJBADPAGE, inconsistent page 11 from storage area DISK1:[TEST5]EMPIDS_OVER.RDA;1 needs AIJ sequence number 0 %RMU-I-AIJBADPAGE, inconsistent page 60 from storage area DISK1:[TEST5]EMPIDS_LOW.RDA;1 needs AIJ sequence number 0 %RMU-I-AIJBADPAGE, inconsistent page 123 from storage area DISK1:[TEST5]EMPIDS_OVER.RDA;1 needs AIJ sequence number 0 %RMU-I-LOGRECDB, recovering database file DISK2:[TEST5]MF_PERSONNEL.RDB;1 %RMU-I-LOGOPNAIJ, opened journal file DISK3:[TEST5]MF_PERSONNEL.AIJ;1 %RMU-I-AIJONEDONE, AIJ file sequence 0 roll-forward operations completed %RMU-I-LOGRECOVR, 1 transaction committed %RMU-I-LOGRECOVR, 0 transactions rolled back %RMU-I-LOGRECOVR, 0 transactions ignored %RMU-I-AIJNOACTIVE, there are no active transactions %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-LOGSUMMARY, total 1 transaction committed %RMU-I-LOGSUMMARY, total 0 transactions rolled back %RMU-I-LOGSUMMARY, total 0 transactions ignored %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJGOODPAGE, page 11 from storage area DISK1:[TEST5]EMPIDS_OVER.RDA;1 is now consistent %RMU-I-AIJGOODPAGE, page 60 from storage area DISK1:[TEST5]EMPIDS_LOW.RDA;1 is now consistent %RMU-I-AIJGOODPAGE, page 123 from storage area DISK1:[TEST5]EMPIDS_OVER.RDA;1 is now consistent %RMU-I-AIJFNLSEQ, to start another AIJ file recovery, the sequence number needed will be 0 Example 5 In the following example, note that the backed up AIJ files are specified in the order B1, B3, B2, B4 representing sequence numbers 1, 3, 2, 4. The /ORDER_AIJ_FILES sorts the journals to be applied into ascending sequence order and then is able to remove B1 from processing because the database recovery starts with AIJ file sequence 2 as shown in the RMU/RESTORE output. $ RMU/RESTORE/NEW/NOCDD/NOAFTER FOO %RMU-I-RESTXT_00, Restored root file DUA0:[DB]FOO.RDB;16 . . . %RMU-I-AIJRECFUL, Recovery of the entire database starts with AIJ file sequence 2 %RMU-I-COMPLETED, RESTORE operation completed at 24-MAY-2007 12:23:32.99 $! $ RMU/RECOVER/LOG/ORDER_AIJ_FILES B1,B3,B2,B4 . . . %RMU-I-LOGOPNAIJ, opened journal file DUA0:[DB]B2.AIJ;24 %RMU-I-LOGRECSTAT, transaction with TSN 0:256 ignored %RMU-I-LOGRECSTAT, transaction with TSN 0:257 ignored %RMU-I-RESTART, restarted recovery after ignoring 2 committed transactions %RMU-I-AIJONEDONE, AIJ file sequence 2 roll-forward operations completed %RMU-I-LOGRECOVR, 0 transactions committed %RMU-I-LOGRECOVR, 0 transactions rolled back %RMU-I-LOGRECOVR, 2 transactions ignored %RMU-I-AIJNOACTIVE, there are no active transactions %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJNXTSEQ, to continue this AIJ file recovery, the sequence number needed will be 3 . . . Example 6 The following example shows the "/CONFIRM=ABORT" syntax used so that RMU/RECOVER will not continue rolling forward if a sequence gap is detected. This is the default behavior if /NOCONFIRM is specified or for batch jobs. Note that the exit status of RMU will be "%RMU-E-AIJRECESQ" if the recovery is aborted due to a sequence gap. It is always a good policy to check the exit status of RMU, especially when executing RMU in batch jobs. RMU/RECOVER/CONFIRM=ABORT/LOG/ROOT=user$test:foo faijbck1,faijbck2,faijbck4 %RMU-I-LOGRECDB, recovering database file DEVICE:[DIRECTORY]FOO.RDB;1 %RMU-I-LOGOPNAIJ, opened journal file DEVICE:[DIRECTORY]FAIJBCK4.AIJ;1 at 25-FEB-2009 17:27:42.29 %RMU-W-AIJSEQAFT, incorrect AIJ file sequence 8 when 7 was expected %RMU-E-AIJRECESQ, AIJ roll-forward operations terminated due to sequence error %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-LOGSUMMARY, total 2 transactions committed %RMU-I-LOGSUMMARY, total 0 transactions rolled back %RMU-I-LOGSUMMARY, total 0 transactions ignored %RMU-I-AIJFNLSEQ, to start another AIJ file recovery, the sequence number needed will be 7 %RMU-I-AIJNOENABLED, after-image journaling has not yet been enabled Example 7 The following example shows the "/CONFIRM=CONTINUE" syntax used so that RMU/RECOVER will continue rolling forward if a sequence gap is detected. RMU/RECOVER/CONFIRM=CONTINUE/LOG/ROOT=user$test:foo faijbck1,faijbck2,faijbck4 %RMU-I-LOGRECDB, recovering database file DEVICE:[DIRECTORY]FOO.RDB;1 %RMU-I-LOGOPNAIJ, opened journal file DEVICE:[DIRECTORY]FAIJBCK4.AIJ;1 at 25-FEB-2009 17:26:04.00 %RMU-W-AIJSEQAFT, incorrect AIJ file sequence 8 when 7 was expected %RMU-I-AIJONEDONE, AIJ file sequence 8 roll-forward operations completed %RMU-I-LOGRECOVR, 1 transaction committed %RMU-I-LOGRECOVR, 0 transactions rolled back %RMU-I-LOGRECOVR, 0 transactions ignored %RMU-I-AIJNOACTIVE, there are no active transactions %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJNXTSEQ, to continue this AIJ file recovery, the sequence number needed will be 9 %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-LOGSUMMARY, total 3 transactions committed %RMU-I-LOGSUMMARY, total 0 transactions rolled back %RMU-I-LOGSUMMARY, total 0 transactions ignored %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJFNLSEQ, to start another AIJ file recovery, the sequence number needed will be 9 %RMU-I-AIJNOENABLED, after-image journaling has not yet been enabled 3 Resolve Recovers a corrupted database and resolves an unresolved distributed transaction by completing the transaction. See the Oracle Rdb7 Guide to Distributed Transactions for complete information on unresolved transactions and for information on the transactions managers (DECdtm and Encina) supported by Oracle Rdb. 4 Description Use the RMU Recover Resolve command to commit or abort any unresolved distributed transactions in the after-image journal (.aij) file. You must complete the unresolved transactions to the same state (COMMIT or ABORT) in every .aij file affected by the unresolved transactions. The RMU Recover Resolve command performs the following tasks: o Displays identification information for an unresolved transaction. o Prompts you for the state to which you want the unresolved transaction resolved (if you did not specify the State qualifier on the command line). If you are using DECdtm to manage the transaction, you can specify COMMIT, ABORT, or IGNORE. If you are using an XA transaction, you can specify COMMIT or ABORT. o Prompts for confirmation of the state you specified o Commits, aborts, or ignores the unresolved transaction o Continues until it displays information for all unresolved transactions 4 Format (B)0RMU/Recover/Resolve aij-file-name Command Qualifiers x Defaults x /Active_IO=max-reads x See the RMU/Recover command /Aij_Buffers=integer x See the RMU/Recover command /Areas[=storage-area[,...]] x See the RMU/Recover command /[No]Confirm x See description /Format={Old_File|New_Tape} x See the RMU/Recover command /Label=(label-name-list) x See the RMU/Recover command /[No]Log x See the RMU/Recover command /[No]Media_Loader x See the RMU/Recover command /[No]Online x See the RMU/Recover command /[No]Rewind x See the RMU/Recover command /Root=root-file-name x See the RMU/Recover command /State=option x See description /[No]Trace x See the RMU/Recover command /Until=date-time x See the RMU/Recover command 4 Parameters 5 aij-file-name The name of the file containing the after-image journal. This cannot be an optimized after-image journal (.oaij) file. The default file extension is .aij. 4 Command_Qualifiers 5 Confirm Confirm Noconfirm Prompts you for confirmation of each transaction state you alter. The default for interactive processing is Confirm. Specify the Noconfirm qualifier to suppress this prompt. The default for batch processing is Noconfirm. 5 State State=option Specifies the state to which all unresolved transactions will be resolved. If you are using DECdtm to manage your distributed transaction, options for the State qualifier are: o Commit-Commits all unresolved transactions. o Abort- Aborts all unresolved transactions. o Ignore-Does not resolve any transactions. If you are using Encina to manage your distributed transaction, options for the State qualifier are: o Commit-Commits all unresolved transactions. o Abort- Aborts all unresolved transactions. If you do not specify the State qualifier, Oracle RMU prompts you to enter an action, for each unresolved transaction in that .aij file. If DECdtm is managing your transaction and you enter Ignore, Oracle RMU-instead of resolving the transaction- attempts to contact the coordinator to resolve the transaction. The transaction remains unresolved until the coordinator becomes available again and instructs the transaction to complete or until you manually complete the transaction by using the RMU Recover Resolve command again. For more information about the activities of the coordinator, see the Oracle Rdb7 Guide to Distributed Transactions. Because a coordinator is not involved with transactions managed by Encina, the Ignore option is not valid for XA transactions. 4 Usage_Notes o To use the RMU Recover Resolve command for a database, you must have the RMU$RESTORE privilege in the root file for the database or the OpenVMS SYSPRV or BYPASS privilege. o If you have restored the database by using the New qualifier and have not deleted the corrupted database, use the Root qualifier to override the original file specification for the database root file. o After it rolls forward from the .aij file specified on the command line, Oracle RMU prompts you for the name of the next .aij file. If there are more .aij files to roll forward, enter the file name, including the version number for that .aij file. If there are no other .aij files, press the Return key. For more information about rolling forward and determining transaction sequence numbers for .aij files, see the Oracle Rdb Guide to Database Maintenance. o Note the following points regarding using Oracle Rdb with the Encina transaction manager: - Only databases that were created under Oracle Rdb V7.0 or higher, or converted to V70 or higher, can participate in XA transactions. - To start a distributed transaction, you must have the DISTRIBTRAN database privilege for all databases involved in the transaction. - Oracle Rdb supports only explicit distributed transactions with Encina. This means that your application must explicitly call the Encina routines to start and end the transactions. 4 Examples Example 1 The following command recovers the mf_personnel database and rolls the database forward from the old .aij file to resolve the unresolved distributed transactions. Because the State qualifier is not specified, Oracle RMU will prompt the user for a state for each unresolved transaction. $ RMU RECOVER/RESOLVE MF_PERSONNEL.AIJ;1 Example 2 This example specifies that all unresolved transactions in the mf_personnel.aij file be committed. $ RMU/RECOVER/RESOLVE/STATE=COMMIT MF_PERSONNEL.AIJ For more examples of the RMU Recover Resolve command, see the Oracle Rdb7 Guide to Distributed Transactions. 2 Repair Corrects several types of database problems. You can use the RMU Repair command to: o Repair all types of space area management (SPAM) page corruptions by reconstructing the SPAM pages in one or more storage areas. o Repair all area bit map (ABM) page format errors. o Repair all page tail errors to the satisfaction of the RMU Verify operation by making sure that every database page is in a logical area and contains the appropriate information for that logical area. o Correct some performance problems that might otherwise have to be corrected by exporting and importing the database. o Set damaged or missing segmented string (LIST OF BYTE VARYING) areas that are stored in write-once areas to null. The repair operation cannot correct corrupted user data, or corrupted indexes; use other commands such as the RMU Restore, the RMU Recover, the SQL IMPORT, or the RMU Load command and delete the affected structures to correct these problems. NOTE Use of the Abm or the Initialize=Tsns qualifier disables after-image journaling. After issuing an RMU Repair command with these qualifiers, back up the database and reenable journaling manually. 3 Description Because RMU Repair cannot correct every type of corruption, or guarantee improved performance, Oracle Corporation recommends that you do not use the RMU Repair command unless you have a backup copy or exported copy of your database. You can return to this backup copy of the database if your repair efforts are ineffective. The RMU Repair command operates off line and not in the context of a transaction, so no records are written to the database's .aij file by RMU Repair, and the repaired database cannot be rolled forward with the RMU Recover command. Oracle Corporation recommends that you make a backup copy of the database after using the RMU Repair command; the repair operation issues a message to this effect. Oracle RMU also issues a warning when you use this command on a database with after-image journaling enabled. 3 Format (B)0 Command Qualifiers x Defaults x /[No]Abm x /Noabm /[No]All_Segments x All segments /Areas [={storage-area-list or *}] x See description /Checksum x See description /[No]Initialize=initialize-options x /Noinitialize /[No]Spams x See description /Tables [=table-list] x All nonsystem tables /Worm_Segments x None 3 Parameters 4 root-file-spec A file specification for the database root file for which you want to repair corruption or improve performance. 3 Command_Qualifiers 4 Abm Abm Noabm Causes the reconstruction of the logical area bit map (ABM) pages for areas specified with the Areas qualifier. After-image journaling is disabled when you specify the Abm qualifier. You must explicitly enable after-image journaling after the RMU Repair command completes if you want journaling enabled. The NoAbm qualifier specifies that ABM pages are not to be reconstructed; this is the default. 4 All_Segments All_Segments Noall_Segments The All_Segments qualifier specifies that RMU Repair should retrieve all segments of a segmented string; the Noall_Segments qualifier specifies that RMU Repair should only retrieve the first segment of a segmented string. Specify the Noall_Segments qualifier if you know that the list storage map for any segmented strings stored on the specified areas might have contained multiple areas. For example, if the storage map was created using the following SQL command, Oracle Rdb would store all the segmented strings on AREA1 until AREA1 became full. If AREA1 became full, Oracle Rdb would continue to write the rest of the segments into AREA2. Suppose AREA2 becomes corrupt. In this case, retrieving the first segment from AREA1 is not sufficient; all segments must be retrieved to determine if part of the segmented string is missing. CREATE STORAGE MAP FOR LIST STORE IN (AREA1, AREA2) FOR (TABLE1) IN RDB$SYSTEM; Specifying the Areas qualifier and the All_Segments qualifier is unnecessary and redundant because specifying the All_Segments qualifier causes RMU Repair to check all storage areas regardless of where the segmented string was stored initially. 4 Areas Areas[={storage-area-list or *}] Specifies the storage areas in the database you want to repair. You can specify storage areas by name or by the area's ID number. By default, all the storage areas in the database are repaired. If you specify more than one storage area, separate the storage area names or ID numbers in the storage-area-list with a comma, and enclose the list within parentheses. 4 Checksum Checksum Reads every page in the database storage areas to verify that the checksum on each page is correct. If the checksum on the page is incorrect, it is replaced with the correct checksum. Use the Areas qualifier to specify which storage areas RMU Repair should check. If you do not specify the Areas qualifier, all pages in all storage areas are checked and updated (if incorrect). This qualifier can be used whether or not users are attached to the database. This qualifier is not valid for use with any other qualifiers except the Areas qualifier. 4 Initialize Initialize=initialize-options Noinitialize Allows you to specify initialization options. If more than one option is specified, separate the options with a comma, and enclose the list of options within parentheses. The following options are available for the Initialize qualifier: o Free_Pages The Initialize=Free_Pages qualifier initializes database pages that do not contain data in the selected storage areas (that have a uniform page format). You can use the Initialize=Free_ Pages qualifier to correct BADPTLARE errors found by the RMU Verify command and also to free pages from a table that has many deleted rows. If you specify the default, the Noinitialize qualifier, no database pages are initialized. Frequently, you will receive one or more RMU-W-ABMBITTERR error messages after you issue the RMU Repair command with the Initialize=Free_Pages qualifier. This occurs because the initialization of pages can create new ABM errors. Correct these errors by issuing the RMU Repair command with the Abm qualifier. (However, note that you cannot specify the Initialize=Free_Pages qualifier and the Abm qualifier on the same command line.) If you ignore the RMU-W-ABMBITTERR error messages, extra I/O operations will be performed (one for each RMU-W-ABMBITTERR error you received) when a database query causes a sequential scan of an entire table. If a table residing in a storage area that has a uniform page format is frequently accessed sequentially, the cost of the sequential access is determined by the number of allocated pages. If the maximum size allocated for the table is much larger than the table's average size, the cost of the sequential access can be excessive. By using the RMU Repair command with the Initialize=Free_Pages qualifier, you can purge the allocated but unused database pages from the table. In some cases, there may be a decrease in performance when you insert new data into the table after using this option. As with all Repair options, you should test the performance of the database after executing the command and be prepared to restore the backup made before executing the Repair command if you find that the command results in decreased performance. The initialization of free pages requires access to the Oracle Rdb system tables. You should not initialize free pages until you know that the RDB$SYSTEM storage area (where the system tables are stored) is not corrupted. o Larea_Parameters=options-file This option specifies an options file (default file extension .opt) that contains a list of logical areas and parameter values that RMU Repair uses to update the area inventory page (AIP) before it builds the space area management (SPAM) pages. The Larea_Parameters options file contains lines in the following format: name [/Areas=name][/Delete][/[No]Thresholds=(n[,n[,n]])[/Length=n][/Type=option] A comment can be appended to the line (an exclamation point (!) is the comment character), and a line can be continued (as in DCL) by ending it with a hyphen (-). The logical area can be specified by name or identification number (ID). The logical area named must be present in the AIP, or an error is generated. The Larea_Parameters options are further described as follows: - Areas=name Restricts this line to the logical area that resides in the specified storage area. The storage area can be specified by name or ID. By default, all logical areas with a matching name are altered independently of the storage area in which they reside. You can specify storage area ID numbers with the Areas qualifier. - Delete Specifies that the logical area should be marked as deleted. You will corrupt your database if you delete a logical area that is referenced by Oracle Rdb metadata. - Length=n The Initialize=Length option specifies the record length to store in the logical area inventory entry. RMU Repair uses this value to calculate SPAM thresholds. When columns are deleted from or added to a table, the record length stored in the logical area inventory entry is not updated. Therefore the search for space needed to store a new record may be inefficient, and the SPAM thresholds will not be set properly. You can solve this problem by first correcting the length in the logical area inventory entry, then generating corrected SPAM pages using the RMU Repair command. See Example 2 in the Examples help entry under this command. - Thresholds=(n [,n [,n]]) NoThresholds This option specifies the logical area SPAM thresholds. This is useful only for logical areas that reside in a storage area with a uniform page format. If thresholds are set, they are ignored in a storage area with a mixed page format. See the Oracle Rdb7 Guide to Database Performance and Tuning for information on setting SPAM thresholds. The Nothresholds option specifies that logical area thresholds be disabled. - Type=keyword By specifying a Type, you can update the on-disk logical area type in the AIP. For databases created prior to Oracle Rdb release 7.0.1, the logical area type information in the AIP is unknown. However, the RMU Show Statistics utility depends on this information to display information on a per-logical-area basis. A logical area is a table, B-tree index, hash index, or any partition of one of these. In order to update the on-disk logical area type in the AIP, specify the type as follows: Type=Table Specifies that the logical area is a data table, such as is created with the SQL CREATE TABLE statement. Type=Btree Specifies that the logical area is a B-tree index, such as is created with the SQL CREATE INDEX TYPE IS SORTED statement. Type=Hash Specifies that the logical area is a hash index, such as is created with the SQL CREATE INDEX TYPE IS HASHED statement. Type=System Specifies that the logical area is a system record that is used to identify hash buckets. Users cannot explicitly create this type of logical area. This type should not be used for the RDB$SYSTEM logical areas. It does not identify system relations. Type=Blob Specifies that the logical area is a BLOB (LIST OF BYTE VARYING) repository. There is no error checking of the type specified for a logical area. The specified type does not affect the collection of statistics, nor does it affect the readying of the affected logical areas. However, an incorrect type will cause incorrect statistics to be reported by the RMU Show Statistics utility. o Only_Larea_Type The Initialize=Only_Larea_Type option specifies that only the logical area type field is to be updated in the area inventory page (AIP). o Snapshots The Snapshots option allows you to create and initialize new snapshot files. In addition, it removes corrupt snapshot area pages from the Corrupt Page Table (CPT). This is much faster than using the RMU Restore command to do the same thing, especially when just one snapshot file is lost and needs to be created again. The default is not to create new files. When you specify the Confirm option with the Initialize=Snapshots option (Initialize=Snapshots=Confirm), you can use the RMU Repair command not only to initialize, but also to optionally rename, move, or change the allocation of snapshot files. These operations might be necessary when a disk with a snapshot file has a hardware problem or is removed in a hardware upgrade, or when a snapshot file has grown too large and you want to truncate it. The Confirm option causes RMU Repair to prompt you for a name and allocation for one or more snapshot files. If you use the Areas qualifier, you can select the snapshot files in the database that you want to modify. If you omit the Areas qualifier, all the snapshot files for the database are initialized and RMU Repair prompts you interactively for an alternative file name and allocation for each snapshot file. By specifying a new file name for a snapshot file, you can change the location of the snapshot file. By specifying a new allocation for a snapshot file, you can truncate a snapshot file or make it larger. o Tsns The Initialize=Tsns option resets the database transaction state. The default is to not alter the transaction state. After-image journaling is disabled when you specify the Initialize=Tsns option. You must explicitly enable after-image journaling after the RMU Repair command completes if you want journaling enabled. This operation is useful when the database transaction sequence number (TSN) approaches the maximum allowable value and the TSN values must be initialized to zero. The TSN value is contained in a quadword with the following decimal format: high longword : low longword The high longword can hold a maximum user value of 32768 (215) and the low longword can hold a maximum user value of 4,294,967,295 (232). A portion of the high-longword is used by Oracle Rdb for overhead. Initialization of the TSN values requires reading and writing to each page of the database, so the Areas qualifier is not meaningful. It also requires initialization of the snapshot areas even if the Snapshots option has not been specified. The Tsns initialization option carries the following restrictions: - It cannot be performed if the Replication Option for Rdb is being used unless all transfers have been completed. RMU Repair will ask for confirmation if an RDB$TRANSFERS table is defined. - Old journal files will not be applicable to this repaired database. After TSNs have been initialized, you must reenable after-image journaling if you want journaling enabled. After the RMU Repair command completes, a full and complete backup operation should be performed on the database as soon as is practical. This operation ensures that new journaled changes can be applied to the restored database in the event that a restore operation should become necessary. 4 Spams Spams Nospams Reconstructs the SPAM pages for the areas you specify with the Areas qualifier. If you specify the Nospams qualifier, the SPAM pages are not reconstructed. The default is the Spam qualifier if you do not specify any of the following qualifiers for the RMU Repair command: o ABM o Initialize=Free_Pages o Initialize=Snapshots o Initialize=Snapshots=Confirm If you use any of these qualifiers, the NoSpam qualifier is the default. When columns are deleted from or added to a table, the record length stored in the logical area inventory entry is not updated. Therefore the search for space needed to store a new record may be inefficient, and the SPAM thresholds will not be set properly. You can solve this problem by first correcting the length in the logical area inventory entry, then generating corrected SPAM pages using the RMU Repair command. See Example 2 in the Examples help entry under this command. 4 Tables Tables[=table-list] Specifies the list of tables that you want RMU Repair to check for complete segmented strings. If no tables are listed, then all nonsystem tables are examined. (System tables do not store their segmented strings in write-once areas.) Note that RMU Repair has no knowledge of which storage areas contain segmented strings from a particular table; thus, the default is to search all tables. 3 Usage_Notes o To use the RMU Repair command for a database, you must have the RMU$ALTER privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o Enable detected asynchronous prefetch to achieve the best performance of this command. Beginning with Oracle Rdb V7.0, by default, detected asynchronous prefetch is enabled. You can determine the setting for your database by issuing the RMU Dump command with the Header qualifier. If detected asynchronous prefetch is disabled, and you do not want to enable it for the database, you can enable it for your RMU Repair operations by defining the following logicals at the process level: $ DEFINE RDM$BIND_DAPF_ENABLED 1 $ DEFINE RDM$BIND_DAPF_DEPTH_BUF_CNT P1 P1 is a value between 10 and 20 percent of the user buffer count. o The Areas qualifier can be used with indirect file references. See the Indirect-Command-Files help entry. o Oracle Corporation recommends that you use the RMU Backup command to perform a full backup operation on your database before using the RMU Repair command on the database. o Use the SQL SHOW STORAGE AREA statement to display the new location of a snapshot (.snp) file and the RMU Dump command with the Header qualifier to display the new allocation. o Be careful when you specify names for new .snp files with the RMU Repair command. If you specify the name of a file that already exists and was created for the database, it will be initialized as you requested. If you mistakenly initialize a live database file in this way, do not use the database until the error is corrected. Use the RMU Restore command to restore the database to the condition it was in when you backed it up just prior to issuing the RMU Repair command. If you did not back up the database before issuing the RMU Repair command, you must restore the database from your most recent backup file and then recover from .aij files (if the database had after-image journaling enabled). If you specify the wrong .snp file (for example, if you specify jobs.snp for all the .snp file name requests in Example 3 in the Examples help entry under this command), you can correct this by issuing the RMU Repair command again with the correct .snp file names. After the RMU Repair command completes, delete old .snp files and use the RMU Backup command to perform a full backup operation on your database. 3 Examples Example 1 The following command repairs SPAM page corruption for all the storage areas in the mf_personnel database. No area bit map (ABM) pages are reconstructed because the Abm qualifier is not specified. $ RMU/REPAIR MF_PERSONNEL Example 2 When columns are deleted from or added to a table, the record length stored in the logical area inventory entry is not updated. Therefore the search for space needed to store a new record may be inefficient, and the SPAM thresholds will not be set properly. You can solve this problem by first correcting the length in the logical area inventory entry, then generating corrected SPAM pages using the RMU Repair command. For example, suppose the Departments table was stored in the departments.rda uniform page format storage area and the Budget_ Projected column (integer data type = 4 bytes) was deleted. As a result of this deletion, the row length changed from 47 bytes to 43 bytes. You can specify a smaller record length (43 bytes) in the fix_departments.opt options file to more efficiently use space in the storage area. $ CREATE FIX_DEPARTMENTS.OPT DEPARTMENTS /LENGTH=43 Then, the following RMU Repair command specifies the record length to store in the logical area inventory entry for this logical area and rebuilds the SPAM pages: $ RMU/REPAIR/SPAMS/INITIALIZE=LAREA_PARAMETERS=FIX_DEPARTMENTS.OPT - _$ MF_PERSONNEL Example 3 The following RMU Repair command initializes and renames departments.snp; initializes and moves salary_history.snp; and initializes, moves, and truncates jobs.snp: $ RMU/REPAIR/NOSPAMS/INITIALIZE=SNAPSHOTS=CONFIRM - _$ /AREAS=(DEPARTMENTS,JOBS,SALARY_HISTORY) MF_PERSONNEL %RMU-I-FULBACREQ, A full backup of this database should be performed after RMU Repair Area DEPARTMENTS snapshot filename [SQL1:[TEST]DEPARTMENTS.SNP;1]: NEW_DEPT Area DEPARTMENTS snapshot file allocation [10]? Area SALARY_HISTORY snapshot filename [SQL1:[TEST]SALARY_HISTORY.SNP;1]: SQL2:[TEST] Area SALARY_HISTORY snapshot file allocation [10]? Area JOBS snapshot filename [SQL1:[TEST]JOBS.SNP;1]: SQL2:[TEST2] Area JOBS snapshot file allocation [10]? 5 Example 4 The following RMU Repair command finds incorrect checksums in the EMPIDS_LOW storage area and updates them to reflect the correct checksum: $ RMU/REPAIR MF_PERSONNEL.RDB/AREA=EMPIDS_LOW/CHECKSUM Example 5 The following command updates an AIP type for a table: $ RMU/REPAIR MF_PERSONNEL /INITIALIZE=LAREA_PARAMETERS=TABLE.OPT Type the TABLE.OPT file to show the contents of the file. $ TYPE TABLE.OPT EMPLOYEES /TYPE=TABLE Example 6 The following command updates an AIP type for a storage area: $ RMU/REPAIR MF_PERSONNEL /INITIALIZE=LAREA_PARAMETERS=AREAS.OPT Type the AREAS.OPT file to show the contents of the file. $ TYPE AREAS.OPT EMPLOYEES /AREA=EMPIDS_OVER /TYPE=TABLE 2 Resolve Resolves all unresolved distributed transactions for the specified database. For more information on unresolved transactions, see the Oracle Rdb7 Guide to Distributed Transactions and the Oracle Rdb Release Notes. 3 Description Use the RMU Resolve command to commit or abort any unresolved distributed transactions in the database. You must resolve the unresolved transactions to the same state (Commit or Abort) in every database affected by the unresolved transactions. RMU Resolve performs the following tasks: o Displays identification information for an unresolved transaction. o Prompts you for the state (Commit or Abort) to which you want the unresolved transaction resolved (if you did not specify the State qualifier on the command line). o Prompts you for confirmation of the state you chose. o Commits or aborts the unresolved transaction. If you commit or abort the unresolved transaction, it is resolved and cannot be resolved again. o Continues to display and prompt for states for subsequent unresolved transactions until it has displayed information for all unresolved transactions. Use the Parent_Node, Process, or Tsn qualifiers to limit the number of unresolved transactions that Oracle RMU displays. Use the Users and State=Blocked qualifiers with the RMU Dump command to determine values for the Parent_Node, Process, and Tsn qualifiers. 3 Format (B)0RMU/Resolve root-file-spec Command Qualifiers x Defaults x /[No]Confirm x See description /[No]Log x Setting of DCL VERIFY flag /Parent_Node=node-nam e x See description /Process=process-id x See description /State=options x None /Tsn=tsn x See description 3 Parameters 4 root-file-spec The database root file for which you want to resolve unresolved transactions. 3 Command_Qualifiers 4 Confirm Confirm Noconfirm Prompts you for confirmation of each unresolved transaction. This is the default for interactive processing. Specify the Noconfirm qualifier to suppress this prompt. This is the default for batch processing. 4 Log Log Nolog Specifies whether the processing of the command is reported to SYS$OUTPUT. Specify the Log qualifier to request that summary information about the resolve operation be reported to SYS$OUTPUT and the Nolog qualifier to prevent this reporting. If you specify neither, the default is the current setting of the DCL VERIFY flag. (The DCL SET VERIFY command controls the setting of the DCL VERIFY flag.) 4 Parent_Node Parent_Node=node-name Specifies the node name to limit the selection of transactions to those originating from the specified node. If you omit the Parent_Node qualifier, RMU Resolve includes transactions originating from all nodes. You cannot specify the Tsn or Process qualifier with the Parent_ Node qualifier. The Parent_Node qualifier is not valid for XA transactions. 4 Process Process=process-id Specifies the process identification to limit the selection of transactions to those associated with the specified process. If you omit this qualifier, RMU Resolve includes all processes with transactions attached to the specified database. You cannot specify the Parent_Node or Tsn qualifier with the Process qualifier. 4 State State=options Specifies the state to which all unresolved transactions be resolved. Options for the State qualifier are: o Commit-Commits unresolved transactions. o Abort-Aborts unresolved transactions. If you do not specify the State qualifier, RMU Resolve prompts you to enter an action, Commit or Abort, for each unresolved transaction on that database. 4 Tsn Tsn=tsn Specifies the transaction sequence number (TSN) of the unresolved transactions whose state you want to modify. The TSN value is contained in a quadword with the following decimal format: high longword : low longword The high longword can hold a maximum user value of 32768 (215) and the low longword can hold a maximum user value of 4,294,967,295 (232). A portion of the high longword is used by Oracle Rdb for overhead. When you specify a TSN, you can omit the high longword and the colon if the TSN fits in the low longword. For example 0:444 and 444 are both valid TSN input values. If you omit the Tsn qualifier, RMU Resolve includes all the unresolved transactions. You cannot specify the Parent_Node or the Process qualifier with the Tsn qualifier. 3 Usage_Notes o To use the RMU Resolve command for a database, you must have the RMU$RESTORE privilege in the root file ACL for the database or the OpenVMS SYSPRV or BYPASS privilege. 3 Examples Example 1 The following command specifies that the first displayed unresolved transaction in the MF_PERSONNEL database be changed to the Abort state and rolled back: $ RMU/RESOLVE/LOG/STATE=ABORT MF_PERSONNEL Example 2 The following command will display a list of all transactions coordinated by node GREEN and might be useful if node GREEN failed while running an application that used the DECdtm two- phase commit protocol: $ RMU/RESOLVE/PARENT_NODE=GREEN MF_PERSONNEL Example 3 The following command displays a list of all transactions initiated by process 41E0364A. The list might be useful for resolving transactions initiated by this process if the process were deleted. $ RMU/RESOLVE/PROCESS=41E0364A MF_PERSONNEL Example 4 The following command completes unresolved transactions for the MF_PERSONNEL database, and confirms and logs the operation: $ RMU/RESOLVE/LOG/CONFIRM MF_PERSONNEL For more examples of the RMU Resolve command, see the Oracle Rdb7 Guide to Distributed Transactions. 2 Restore Restores a database to the condition it was in at the time a full or incremental backup operation was performed with an RMU Backup command. In addition, if after-image journal (.aij) files have been retained, RMU Restore attempts to apply any pre- existing .aij files to recover the database completely. See the Description help entry under this command for details on the conditions under which RMU Restore attempts an automatic .aij file recovery as part of the restore operation. When you use the RMU Restore command to restore the database to a system with a more recent version of Oracle Rdb software, an RMU Convert command with the Noconfirm and Commit qualifiers is automatically executed as part of RMU Restore. Therefore, by executing the RMU Restore command, you convert that database to the current version. See the Oracle Rdb Installation and Configuration Guide for the proper backup procedure prior to installing a new release of Oracle Rdb and restoring (or converting) databases. When you use the RMU Restore command to restore a database that was recently RMU/Converted but with the /NoCommit qualifier, the behavior is different than that stated above. /Commit is the default for an RMU Restore of an uncommited database (a database that contains both current and previous versions of the metadata that was converted by specifying RMU/CONVERT/NOCOMMIT or RMU/RESTORE/NOCOMMIT) but ONLY if the noncommited database being restored is NOT of the current Rdb version. RMU/RESTORE/COMMIT and RMU/RESTORE/NOCOMMIT only take effect if RMU/RESTORE needs to call RMU/CONVERT because the database being restored is of a previous Rdb version. If the /COMMIT is specified or defaulted for the Restore of a database of the current level, it is ignored. In this case, an RMU/CONVERT/COMMIT must be used to commit the previous uncommited restore or conversion. NOTE When you restore a database, default or propagated OpenVMS access control entries (ACEs) for the database root (.rdb) file take precedence over any Oracle RMU database access you might have. Therefore, if default or propagated entries are in use, you must use the RMU Show Privilege and RMU Set Privilege commands after a restore operation completes to verify and correct the Oracle RMU access. (You can tell if default or propagated entries are in use because RMU Restore displays the warning message "RMU-W-PREVACL, Restoring the root ACL over a pre-existing ACL". This is a normal condition if the RMU Restore command was invoked from the CDO utility.) To use RMU Show Privilege and RMU Set Privilege commands, you must have the rights to edit the access control list (ACL) using RMU$SECURITY access (which is VMS BIT_ 15 access in the access control entry (ACE)) and also (READ+WRITE+CONTROL) access. (Note that you can grant yourself BIT_15 access by using the DCL SET ACL command if you have (READ+WRITE+CONTROL) access. If you do not have the required access after a restore operation to make the needed changes, someone with the required access or OpenVMS BYPASS or SECURITY access must examine and correct the ACL. This behavior exists in Oracle RMU to prevent someone from using Oracle RMU to override the existing OpenVMS security policy. 3 Description RMU Restore rebuilds a database from a backup file, produced earlier by an RMU Backup command, to the condition the database was in when the backup operation was performed and attempts to automatically recover the .aij files to provide a fully restored and recovered database. You can specify only one backup file parameter in an RMU Restore command. If this parameter is a full backup file, you cannot use the Incremental qualifier. However, you must use the Incremental qualifier if the parameter names an incremental backup file. RMU Restore attempts automatic .aij file recovery by default when you issue a database restore command if you are using fixed- size .aij files, if .aij files have been retained, and if a database conversion has not been performed. (The .aij files are not retained when you specify any of the following qualifiers: Aij_Options, After_Journal, or Duplicate.) RMU Restore does not attempt automatic .aij file recovery if you have backed up any of your .aij files (using the RMU Backup After_Journal command) because RMU Restore has no knowledge of those backup files. In addition, success of the automatic .aij file recovery operation requires that the following criteria be met: o Fixed-size after-image journaling is in effect. o The .aij files must be on disk (not on tape). o The .aij files must not have been marked as inaccessible at the time the database backup operation was performed. o The .aij files must exist and have proper privileges for both read and write operations. o The .aij files must be able to be accessed exclusively; failure indicates that an .aij file is in use by another database user. o The .aij files must have a nonzero length. o The .aij files must have valid header information that corresponds to the current Oracle Rdb product and version number. o The sequence number in the .aij file header must not conflict with the restored definition in the database root information. o The original .rdb file name must not exist. NOTE RMU Restore attempts automatic .aij file recovery when you restore a database from a full, incremental, by-area, or by-page backup file. However, in some cases, you will want to disable this feature by using the Norecovery qualifier. Specifically, you should specify the Norecovery qualifier if either of the following are true: o You are restoring the database from a previous version of Oracle Rdb. o You need to issue more than one RMU Restore command to completely restore the database. For example, if you intend to restore a database by first issuing a full RMU Restore command followed by the application of one or more RMU Restore commands with the Incremental or Area qualifiers, you must specify the Norecovery qualifier on all but the last RMU Restore command in the series you intend to issue. Allowing Oracle RMU to attempt automatic recovery with a full restore operation when you intend to apply additional incremental, by-area, or by-page backup files can result in a corrupt database. RMU Restore does not attempt automatic .aij file recovery if any of the following conditions are true: o The database has been converted since the time you created the backup file that you are attempting to restore. o The first .aij file is not available (perhaps because it has been backed up). o After-image journaling was disabled when the backup operation was performed. o After-image journaling was disabled when the database (or portion of it) was lost. o You specify the Aij_Options, After_Journal, or Duplicate qualifier with the RMU Restore command. If RMU Restore attempts automatic .aij file recovery but fails, you can still recover your database by using the RMU Recover command if the restore operation was successful. NOTE Using the DCL COPY command with a multifile database (assuming the files are copied to a new location) will result in an unsupported, unusable database. This happens because the DCL COPY command cannot update the full file specification pointers (stored in the database root file) to the other database files (.rda, .snp, and optional .aij). You can rename or move the files that comprise a multifile Oracle Rdb database by using one of the following commands: o The RMU Backup and RMU Restore commands o The SQL EXPORT and IMPORT statements o The RMU Move_Area command o The RMU Copy_Database command By default, RMU Restore integrates the metadata stored in the database root (.rdb) file with the data dictionary copy of the metadata (assuming the data dictionary is installed on your system). However, you can prevent dictionary integration by specifying the Nocdd_Integrate qualifier. When you specify the Incremental or Area qualifiers, do not specify the following additional qualifiers: Directory Nodes_Max New_Version Nonew_Version Users_Max The RMU Restore command ignores the Confirm qualifier if you omit the Incremental qualifier. Also, you must specify the Root qualifier when you restore an incremental backup file to a new version of the database, renamed database, or a restored database in a new location. See the Usage Notes subentry for information on restoring a database from tape. 3 Format (B)0RMU/Restore backup-file-spec [storage-area-name[,...]] Command Qualifiers x Defaults x /[No]Acl x /Acl /Active_IO=max-reads x /Active_IO=3 /[No]After_Journal=file-spec x See description /[No]Aij_Options=journal-opts x See description /Area x See description /[No]Cdd_Integrate x /Cdd_Integrate /Close_Wait=n x See description /[No]Commit x /Commit /[No]Confirm x See description /Directory=directory-spec x See description /Disk_File[=(Reader_Threads=n)] x /Disk_File=(Reader_Threads=1) /[No]Duplicate x /Noduplicate /Encrypt=({Value=|Name=}[,Algorithm=]) x See description /Global_Buffers=global-buffer-options x Current value /Incremental x Full restore /Journal=file-name x See description /Just_Corrupt x See description /Label=(label-name-list) x See description (B)0/Librarian[=options] x None /Loader_Synchronization x See description /Local_Buffers=local-buffer-options x Current value /[No]Log[=Brief|Full] x Current DCL verify value /Master x See description /[No]Media_Loader x See Description /[No]New_Version x /Nonew_Version /Nodes_Max=number-cluster-nodes x See description /[No]Online x /Noonline /Open_Mode={Automatic|Manual} x Current value /Options=file-spec x None /Page_Buffers=number-buffers x /Page_Buffers=3 /Path=cdd-path x Existing value /Prompt={Automatic|Operator|Client} x See description /[No]Recovery[=Aij_Buffers=n] x See description /[No]Rewind x /Norewind /Root=root-file-spec x Existing value /Transaction_Mode=(mode-list) x /Transaction_Mode=Current /Users_max=number-users x Existing value /Volumes=n x /Volumes=1 (B)0File or Area Qualifiers x Defaults x /Blocks_Per_Page=integer x See description /Extension= {Disable|Enable} x Current value /File=file-spec x See description /Just_Corrupt x See description /Read_Only x Current value /Read_Write x Current value /Snapshot=(Allocation=n,File=file-spec) x See description /[No]Spams x Current value /Thresholds=(val1[,val2[,val3]]) x Current value 3 Parameters 4 backup-file-spec A file specification for the backup file produced by a previous RMU Backup command. Note that you cannot perform a remote restore operation on an .rbf file that has been backed up to tape and then copied to disk. The default file extension is .rbf. Depending on whether you are performing a restore operation from magnetic tape, disk, or multiple disks, the backup file specification should be specified as follows: o To restore from magnetic tape: If you used multiple tape drives to create the backup file, the backup-file-spec parameter must be provided with (and only with) the first tape drive name. Additional tape drive names must be separated from the first and subsequent tape drive names with commas, as shown in the following example: $ RMU/RESTORE /REWIND $111$MUA0:PERS_FULL_NOV30.RBF,$112$MUA1: o To restore from single or multiple disk files: If you used multiple disk files to create the backup file, the backup-file-spec parameter must be provided with (and only with) the first disk device name. Additional disk device names must be separated from the first and subsequent disk device names with commas. You must also be sure to include the Disk_ File qualifier. For example: $ RMU/RESTORE/DISK_FILE DISK1:[DIR1]MFP.RBF,DISK2:[DIR2],DISK3:[DIR3] As an alternative to listing the disk device names on the command line (which, if you use several devices, can exceed the line-limit length for a command line), you can specify an options file in place of the backup-file-spec. For example: $ RMU/RESTORE/DISK_FILE "@DEVICES.OPT" The contents of devices.opt might appear as follows: DISK1:[DIR1]MFP.RBF DISK2:[DIR2] DISK3:[DIR3] The backup files referenced from such an options file are: DISK1:[DIR1]MFP.RBF DISK2:[DIR2]MFP01.RBF DISK3:[DIR3]MFP02.RBF 4 storage-area-name storage-area-name[,...] A storage area name from the database. This parameter is optional. Use it in the following situations: o When you want to change the values for thresholds or blocks per page. o When you want to change the names specified with the Snapshot or the File qualifier for the restored database. o If you want to restore only selected storage areas from your backup file, you must use the Area qualifier and specify the names of the storage areas you want to restore in either the storage-area-name parameter in the RMU Restore command line, or in the file specified with the Options qualifier. To use this option, specify the storage area name rather than the file specification for the storage area. By using the RMU Backup and RMU Restore commands, you can back up and restore selected storage areas of your database. This Oracle RMU backup and restore by-area feature is designed to: o Speed recovery when corruption occurs in some (not all) of the storage areas of your database. o Reduce the time needed to perform backup operations because some data (data in read-only storage areas, for example) does not need to be backed up with every backup operation performed on the database. If you plan to use the RMU Backup and RMU Restore commands to back up and restore only selected storage areas for a database, you must perform full and complete backup operations on the database at regular intervals. A full and complete backup is a full backup (not an incremental backup) operation on all the storage areas in the database. If the database root (.rdb) file is corrupted, you can only recover storage areas up to (but not past) the date of the last full and complete backup operation. Therefore, Oracle Corporation recommends that you perform full and complete backup operations regularly. If you plan to back up and restore only selected storage areas for a database, Oracle Corporation strongly recommends that you enable after-image journaling for the database (in addition to performing the full and complete backup operation on the database as described earlier). That is, if you are not backing up and restoring all the storage areas in your database, you should have after-image journaling enabled. This ensures that you can recover all the storage areas in your database in the event of a system failure. If you do not have after-image journaling enabled and one or more of the areas restored by RMU Restore are not current with the storage areas not restored, Oracle Rdb will not allow any transactions to use the storage areas that are not current in the restored database. In this situation, you can return to a working database by restoring the database, using the backup file from the last full and complete backup operation on the database storage areas. However, any changes made to the database since the last full and complete backup operation was performed are not recoverable. If you have after-image journaling enabled, use the RMU Recover command to apply transactions from the .aij file to storage areas that are not current after the RMU Restore command completes. When the RMU Recover command completes, your database will be consistent and usable. 3 Command_Qualifiers 4 Acl Acl Noacl Allows you to specify whether to restore the root file access control list (ACL) that was backed up. If you specify the Acl qualifier, the root file ACL that was backed up is restored with the database. If the root file ACL was not backed up and you specify the Acl qualifier with the RMU Restore command, then RMU Restore restores the database without a root file ACL. If you specify the Noacl qualifier, the root file ACL is not restored with the database. The default is the Acl qualifier. 4 Active_IO Active_IO=max-reads Specifies the maximum number of read operations from the backup file that RMU Restore attempts simultaneously. The value of the Active_IO qualifier can range from 1 to 5. The default value is 3. Values larger than 3 might improve performance with multiple tape drives. 4 After_Journal After_Journal=file-spec Noafter_Journal NOTE This qualifier is maintained for compatibility with versions of Oracle Rdb prior to Version 6.0. You might find it more useful to specify the Aij_Options qualifier, unless you are interested in creating an extensible .aij file only. (An extensible .aij file is one that is extended by a specified amount when it reaches a certain threshold of fullness- assuming there is sufficient space on the disk where it resides.) Specifies how RMU Restore is to handle after-image journaling and .aij file creation, using the following rules: o If you specify the After_Journal qualifier and provide a file specification, the RMU process creates a new extensible .aij file and enables journaling. o If you specify the After_Journal qualifier but you do not provide a file specification, RMU Restore creates a new extensible .aij file with the same name as the journal that was active at the time of the backup operation. o If you specify the Noafter_Journal qualifier, RMU Restore disables after-image journaling and does not create a new .aij file. Note that if you specify the Noafter_Journal qualifier there will be a gap in the sequence of the .aij files. For example, suppose your database has .aij file sequence number 1 when you back it up. If you issue an RMU Restore command with the Noafter_Journal qualifier, the .aij file sequence number will be changed to 2. This means that you cannot (and do not want to) apply the original .aij file to the restored database (doing so would result in a sequence mismatch). o If you do not specify an After_Journal, Noafter_Journal, Aij_ Options, or Noaij_Options qualifier, RMU Restore recovers the journal state (enabled or disabled) and tries to reuse the .aij file or files. (See the Description help entry under this command for details on when automatic .aij file recovery is not attempted.) When you specify an .aij file name, you should specify a new device and directory for the .aij file. If you do not specify a device and directory, you receive a warning message. To protect yourself against media failures, put the .aij file on a different device from that of your database files. If the original database is lost or corrupted but the journal files are unaffected, you would typically restore the database without the use of either the Aij_Options or the After_Journal qualifier. The After_Journal qualifier conflicts with the Area and Incremental qualifiers; you cannot specify the After_Journal qualifier and either of these two other qualifiers in the same RMU Restore command line. You cannot use the After_Journal qualifier to create fixed-size .aij files; use the Aij_Options qualifier. 4 Aij_Options Aij_Options=journal-opts Noaij_Options Specifies how RMU Restore is to handle after-image journaling and .aij file creation, using the following rules: o If you specify the Aij_Options qualifier and provide a journal-opts file, RMU Restore creates the .aij file or files you specify for the restored database. If only one .aij file is created for the restored database, it will be an extensible .aij file. If two or more .aij files are created for the restored database, they will be fixed-size .aij files (as long as at least two .aij files are always available). Depending on what is specified in the options file, after-image journaling can either be disabled or enabled. o If you specify the Aij_Options qualifier, but do not provide a journal-opts file, RMU Restore disables journaling and does not create any new .aij files. o If you specify the Noaij_Options qualifier, RMU Restore reuses the original .aij file configuration and recovers the journaling state (enabled or disabled) from the backed-up .aij file. o If you do not specify an After_Journal, Noafter_Journal, Aij_ Options, or Noaij_Options qualifier, RMU Restore recovers the journaling state (enabled or disabled) and tries to reuse the .aij file or files. (This is the same as specifying the Noaij_ Options qualifier.) See the Description help entry under this command for details on when automatic .aij file recovery is not attempted. The Aij_Options qualifier conflicts with the Area and Incremental qualifiers; you cannot specify the Aij_Options qualifier and either of these two other qualifiers in the same RMU Restore command line. If the original database is lost or corrupted but the journal files are unaffected, you would typically restore the database without the use of either the Aij_Options or the After_Journal qualifier. See Show After_Journal for information on the format of a journal-opts-file. 4 Area Area Specifies that only the storage areas listed in the storage-area- name parameter on the command line or in the Options file are to be restored. You can use this qualifier to simplify physical restructuring of a large database. By default, the Area qualifier is not specified. When the Area qualifier is not specified, all the storage areas and the database root (.rdb) file are restored. Therefore, if you want to restore all the storage areas, omit the Area qualifier. If you specify the Area qualifier, a valid database root must exist. (First issue the RMU Restore Only Root command with a full backup file to create a valid database if one does not exist.) By using the RMU Backup and RMU Restore commands, you can back up and restore selected storage areas of your database. This Oracle RMU backup- and restore-by-area feature is designed to: o Speed recovery when corruption occurs in some (not all) of the storage areas of your database. o Reduce the time needed to perform backup operations because some data (data in read-only storage areas, for example) does not need to be backed up with every backup operation performed on the database. NOTE When you perform a by-area restore operation, an area may be marked as inconsistent; that is, the area may not be at the same transaction state as the database root when the restore operation completes. This may happen, for example, when automatic aij recovery is disabled with the Norecovery qualifier, or if automatic recovery fails. You can check to see if an area is consistent by using the RMU Show Corrupt_Pages command. If you find that one or more areas are inconsistent, use the RMU Recover command to apply the .aij files. If the .aij files are not available, refer to the section on Clearing an Inconsistent Flag in the Oracle Rdb Guide to Database Maintenance for information on the implications of setting a corrupt area to consistent. Then refer to Set Corrupt_Pages for information on using the Set Corrupt_Pages command to clear the inconsistent flag. If you attempt to restore a database area that is not in the backup file, you receive an error message and, typically, the database will be inconsistent or unusable until the affected area is properly restored. In the following example, the DEPARTMENTS storage area is excluded from the backup operation; therefore, a warning message is displayed when the attempt is made to restore DEPARTMENTS, which is not in the backup file. Note that when this restore operation is attempted on a usable database, it completes, but the DEPARTMENTS storage area is now inconsistent. $ RMU/BACKUP /EXCLUDE=DEPARTMENTS MF_PERSONNEL.RDB - _$ PERS_BACKUP5JAN88.RBF $ RMU/RESTORE /NEW_VERSION /AREA PERS_BACKUP5JAN88.RBF DEPARTMENTS %RMU-W-AREAEXCL, The backup does not contain the storage area - DEPARTMENTS If you create a backup file by using the RMU Backup command and the Exclude qualifier, it is your responsibility to ensure that all areas of a database are restored and recovered when you use the RMU Restore and RMU Recover commands to duplicate the database. The Area qualifier conflicts with the After_Journal and Aij_ Options qualifiers. 4 Cdd_Integrate Cdd_Integrate Nocdd_Integrate Integrates the metadata from the database root (.rdb) file into the data dictionary (assuming the data dictionary is installed on your system). If you specify the Nocdd_Integrate qualifier, no integration occurs during the restore operation. You might want to delay integration of the database metadata with the data dictionary until after the restore operation finishes successfully. You can use the Nocdd_Integrate qualifier even if the DICTIONARY IS REQUIRED clause was used when the database was defined. The Cdd_Integrate qualifier integrates definitions in one direction only-from the database file to the dictionary. The Cdd_Integrate qualifier does not integrate definitions from the dictionary to the database file. 4 Close_Wait Close_Wait=n Specifies a wait time of n minutes before RMU Restore automatically closes the database. You must supply a value for n. In order to use this qualifier, the Open_Mode qualifier on the RMU Restore command line must be set to Automatic. 4 Commit Commit NoCommit Instructs Oracle RMU to commit the converted database to the current version of Oracle Rdb before completing the restore operation. Use this qualifier only when the backup file being restored is from a previous version of Oracle Rdb. The conversion is permanent and the database cannot be returned to the previous version. The NoCommit qualifier instructs Oracle RMU not to commit the converted database. In this case, you can rollback the database to its original version using the RMU Convert command with the Rollback qualifier, or you can permanently commit it to the current version by issuing the RMU Convert command with the Commit qualifier. It is important to either Commit or Rollback the conversion after you have verified that the conversion was successful otherwise unnecessary space is taken up in the database to store the obsolete alternate version of the metadata. (RMU will not let you convert to a newer version if the previous Convert was never committed, even if it was years ago.) The Commit qualifier is the default. 4 Confirm Confirm Noconfirm Specifies that RMU Restore notify you of the name of the database on which you are performing the incremental restore operation. You can thus be sure that you have specified the correct .rdb file name to which the incremental backup file will be applied. This is the default for interactive processing. Confirmation is especially important on an incremental restore operation if you have changed the .rdb file name or created a new version of the database during a restore operation from the full backup file. (You must specify the Root qualifier also to create new version or change the .rdb file name.) Specify the Noconfirm qualifier to have RMU Restore apply the incremental backup file to the database without prompting for confirmation. This is the default for batch processing. RMU Restore ignores the Confirm and Noconfirm qualifiers unless you use the Incremental qualifier. 4 Directory Directory=directory-spec Specifies the default destination for the restored database files. If you specify a file name or file extension, all restored files are given that file name or file extension. There is no default directory specification for this qualifier. If you do not specify the Directory qualifier, RMU Restore attempts to restore all the database files to the directories they were in at the time the backup file was created; if those directories no longer exist, the restore operation fails. See the Usage Notes for information on how this qualifier interacts with the Root and File qualifiers and for warnings regarding restoring database files into a directory owned by a resource identifier. 4 Disk_File Disk_File[=(Reader_Threads=integer)] Specifies that you want to perform a multithreaded restore operation from disk files, floppy disks, or other disks external to the PC. This qualifier must have been specified on the RMU Backup command when the backup files from which you are restoring were created. The Reader_Threads keyword specifies the number of threads that Oracle RMU should use when performing a multithreaded restore operation from disk files. You can specify no more than one reader thread per device specified on the command line (or in the command parameter options file). By default, one reader thread is used. This qualifier and all qualifiers that control tape operations (Label, Loader_Synchronization, Master, Media_Loader, and Rewind) are mutually exclusive. 4 Duplicate Duplicate Noduplicate Specifies a new database with the same content but different identity from that of the original database. The default is the Noduplicate qualifier. The Duplicate qualifier creates a copy of your database that is not expected to remain in sequence with the original database. Note that you cannot interchange after-image journal (.aij) files between the original and duplicate copy of the database because each database is unique. You can create a duplicate database when you use the Duplicate qualifier or create the original database again when you use the Noduplicate qualifier. The Duplicate qualifier conflicts with the Incremental, Area, and Online qualifiers. 4 Encrypt Encrypt=({Value=|Name=}[,Algorithm=]) The Encrypt qualifier decrypts the save set file of a database backup. Specify a key value as a string or, the name of a predefined key. If no algorithm name is specified the default is DESCBC. For details on the Value, Name and Algorithm parameters see HELP ENCRYPT. This feature requires the OpenVMS Encrypt product to be installed and licensed on this system. 4 Global_Buffers Global_Buffers=global-buffer-options Allows you to change the default global buffer parameters when you restore a database. The following options are available: o Disabled Use this option to disable global buffering for the database being restored. o Enabled Use this option to enable global buffering for the database being restored. You cannot specify both the Global_ Buffers=Disabled and Global_Buffers=Enabled qualifiers in the same RMU Restore command. o Total=total-buffers Use this option to specify the number of buffers available for all users. The minimum value you can specify is 2; the maximum value you can specify is the global buffer count stored in the .rdb file. o User_Limit=buffers-per-user Use this option to specify the maximum number of buffers available to each user. If you do not specify a Global_Buffers qualifier, the database is restored with the values that were in effect when the database was backed up. When you specify two or more options with the Global_Buffers qualifier, use a comma to separate each option and enclose the list of options within parentheses. 4 Incremental The Incremental qualifier restores a database from an incremental backup file. Use the Incremental qualifier only when you have first issued an RMU Restore command that names the full backup file that was the basis for this incremental backup file. Each incremental backup file is tied to a particular full backup file. After restoring both the full and the incremental backup files, you have restored the database to the condition it was in when you performed the incremental database backup operation. By default, RMU Restore performs a full restore operation on the backup file. You cannot specify the After_Journal or Just_Corrupt qualifier with the Incremental qualifier. 4 Journal Journal=file-name Allows you to specify a journal file to be used to improve tape performance by a restore operation (including a by-area or just- corrupt restore operation). The backup operation creates the journal file and writes to it a description of the backup operation. This description contains identification of the tape drives, the tape volumes and their contents. The Journal qualifier directs RMU Restore to read the journal file and select only the useful tape volumes. The journal file must be the one created at the time the backup operation was performed. If the wrong journal file is supplied, RMU Restore returns an informational message and does not use the specified journal file to select the volumes to be processed. If you omit the Label qualifier, the restore operation creates a list of volume labels from the contents of the journal file. A by-area restore operation also constructs a list of useful tape volume labels from the journal file; only those volumes are mounted and processed. 4 Label Label=(label-name-list) Specifies the 1- to 6-character string with which the volumes of the backup file have been labeled. The Label qualifier is applicable only to tape volumes. You must specify one or more label names when you use the Label qualifier. You can specify a list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas, and enclose the list of names within parentheses. In a normal restore operation, the Label qualifier you specify with the RMU Restore command should be the same Label qualifier you specified with the RMU Backup command that backed up your database. You can use the Label qualifier with indirect file references. See the Indirect-Command-Files help entry for more information. 4 Librarian Librarian=options Use the Librarian qualifier to restore files from data archiving software applications that support the Oracle Media Management interface. The file name specified on the command line identifies the stream of data to be retrieved from the Librarian utility. If you supply a device specification or a version number it will be ignored. Oracle RMU supports retrieval using the Librarian qualifier only for data that has been previously stored by Oracle RMU using the Librarian qualifer. The Librarian qualifier accepts the following options: o Reader_Threads=n Use the Reader_Threads option to specify the number of backup data streams to read from the Librarian utility. The value of n can be from 1 to 99. The default is one reader thread. The streams are named BACKUP_FILENAME.EXT, BACKUP_FILENAME.EXT02, BACKUP_FILENAME.EXT03, up to BACKUP_FILENAME.EXT99. BACKUP_ FILENAME.EXT is the backup file name specified in the RMU Backup command. The number of reader threads specified for a database restore from the Librarian utility should be equal to or less than the number of writer threads specified for the database backup. If the number of reader threads exceeds the number of writer threads, the number of reader threads is set by Oracle RMU to be equal to the number of data streams actually stored in the Librarian utility by the backup. If the number of reader threads specified for the restore is less than the number of writer threads specified for the backup, Oracle RMU will partition the data streams among the specified reader threads so that all data streams representing the database are restored. The Volumes qualifier cannot be used with the Librarian qualifer. Oracle RMU sets the volume number to be the actual number of data streams stored in the specified Librarian utility. o Trace_file=file-specification The Librarian utility writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian utility. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian utility. The pre-determined values are : - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical_name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian utility can use to specify catalogs or archives where Oracle Rdb backup files are stored, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of any Oracle RMU command that accesses the Librarian application. The following OpenVMS logical names must be defined for use with a Librarian utility before you execute an Oracle RMU backup or restore operation. Do not use the Logical_Names option provided with the Librarian qualifier to define these logical names. o RMU$LIBRARIAN_PATH This logical name must be defined so that the shareable Librarian image can be loaded and called by Oracle RMU backup and restore operations. The translation must include the file type (for example, .exe), and must not include a version number. The shareable Librarian image must be an installed (known) image. See the Librarian utility documentation for the name and location of this image and how it should be installed. For a parallel RMU backup, define RMU$LIBRARIAN_ PATH as a system-wide logical name so that the multiple processes created by a parallel backup can all translate the logical. $ DEFINE /SYSTEM /EXECUTIVE_MODE - _$ RMU$LIBRARIAN_PATH librarian_shareable_image.exe o RMU$DEBUG_SBT This logical name is not required. If it is defined, Oracle RMU will display debug tracing information messages from modules that make calls to the Librarian shareable image. For a parallel RMU backup, the RMU$DEBUG_SBT logical should be defined as a system logical so that the multiple processes created by a parallel backup can all translate the logical. The following lines are from a backup plan file created by the RMU Backup/Parallel/Librarian command: Backup File = MF_PERSONNEL.RBF Style = Librarian Librarian_trace_level = # Librarian_logical_names = (- logical_name_1=equivalence_value_1, - logical_name_2=equivalence_value_2) Writer_threads = # The "Style = Librarian" entry specifies that the backup is going to a Librarian utility. The "Librarian_logical_names" entry is a list of logical names and their equivalence values. This is an optional parameter provided so that any logical names used by a particular Librarian utility can be defined as process logical names before the backup or restore operation begins. For example, some Librarian utilities provide support for logical names for specifying catalogs or debugging. You cannot use device specific qualifiers such as Rewind, Density, or Label with the Librarian qualifier because the Librarian utility handles the storage meda, not Oracle RMU. 4 Loader_Synchronization Loader_Synchronization Allows you to preload tapes in order to minimize the need for operator support. When you specify the Loader_Synchronization qualifier and specify multiple tape drives, the restore operation reads from the first set of tape volumes concurrently, then waits until all concurrent tape operations conclude before assigning the next set of tape volumes. This ensures that the tapes can be loaded into the loaders or stackers in the order required by the restore operation. The Loader_Synchronization qualifier does result in reduced performance. For maximal performance, no drive should remain idle, and the next identified volume should be placed on the first drive that becomes idle. However, because the order in which the drives become idle depends on many uncontrollable factors and cannot be predetermined, the drives cannot be preloaded with tapes. Because the cost of using the Loader_Synchronization qualifier is dependent on the hardware configuration and the system load, the cost is unpredictable. A 5% to 20% additional elapsed time for the operation is typical. You must determine whether the benefit of a lower level of operator support compensates for the loss of performance. The Loader_Synchronization qualifier is most useful for large restore operations. The Loader_Synchronization qualifier has no effect unless you specify the Volumes qualifier also. 4 Local_Buffers Local_Buffers=local-buffer-options Allows you to change the default local buffer parameters when you restore a database. The following options are available: o Number=number-buffers Use this option to specify the number of local buffers available for all users. You must specify a number between 2 and 32,767 for the number-buffers parameter. o Size=buffer-blocks The size (in blocks) for each buffer. You must specify a number between 2 and 64 for the buffer-blocks parameter. If you specify a value smaller than the size of the largest page defined, RMU Restore automatically adjusts the size of the buffer to hold the largest page defined. For example, if you specify the Local_Buffers=Size=8 qualifier and the largest page size for the storage areas in your database is 64 blocks, RMU Restore automatically interprets the Local_Buffers=Size=8 qualifier as though it were a Local_Buffers=Size=64 qualifier. The value you specify for the Size option determines the number of blocks for each buffer, regardless of whether local buffering or global buffering is enabled for the database. If you do not specify a Local_Buffers qualifier, the database is restored with the values that were in effect when the database was backed up. 4 Log Log Log=Brief Log=Full Nolog Specifies whether the processing of the command is reported to SYS$OUTPUT. Specify the Log qualifier to request that the progress of the restore operation be written to SYS$OUTPUT, or the Nolog qualifier to suppress this report. If you specify the Log=Brief option, which is the default if you use the Log option without a qualifier, the log contains the start and completion time of each storage area. If you specify the Log=Full option, the log also contains thread assignment and storage area statistics messages. If you do not specify the Log or the Nolog qualifier, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 4 Master Master Allows you to explicitly state how drives should be used when they are to be accessed concurrently. This is a positional qualifier that designates a tape drive as a master tape drive. When the Master qualifier is used, it must be used on the first drive specified. All additional drives become slaves to that master until the end of the command line, or until the next Master qualifier, whichever comes first. If the Master qualifier is used on a drive that does not have an independent I/O path (not a hardware master), performance decreases. If the Master qualifier is not used, and concurrent tape access is requested (using the Volumes=n qualifier), RMU Restore uses the same automatic configuration procedure it employs with the backup operation to select the master drives. Using the Master qualifier is an error if you do not specify concurrent tape access (you do not specify the Volumes=n qualifier). See the description of the Volumes qualifier for further information on specifying concurrent tape access. 4 Media_Loader Media_Loader Nomedia_Loader Use the Media_Loader qualifier to specify that the tape device from which RMU Restore is reading the backup file has a loader or stacker. Use the Nomedia_Loader qualifier to specify that the tape device does not have a loader or stacker. By default, if a tape device has a loader or stacker, RMU Restore should recognize this fact. However, occasionally RMU Restore does not recognize that a tape device has a loader or stacker. Therefore, after reading the first tape, RMU Restore issues a request to the operator for the next tape, instead of requesting the next tape from the loader or stacker. Similarly, sometimes RMU Restore behaves as though a tape device has a loader or stacker when actually it does not. If you find that RMU Restore is not recognizing that your tape device has a loader or stacker, specify the Media_Loader qualifier. If you find that RMU Restore expects a loader or stacker when it should not, specify the Nomedia_Loader qualifier. 4 New_Version New_Version Nonew_Version Specifies whether new versions of database files should be produced if the destination device and directory contain a previous version of the database files. If you use the New_Version qualifier, the new database file versions are produced. The New_Version qualifier conflicts with the Incremental qualifier. If you use the Nonew_Version qualifier, the default, an error occurs if an old copy exists of any of the database files being restored. A restore operation that creates a new database root (.rdb) file must always either disable after-image journaling or create a new .aij file. Attempting to use a pre-existing .aij file with a restored database corrupts the journal and makes future recovery from .aij files impossible. The New_Version qualifier cannot and does not apply to the .aij file. 4 Nodes_Max Nodes_Max=number-cluster-nodes Specifies a new upper limit on the number of VMScluster nodes from which users can access the restored database. The Nodes_Max qualifier accepts values between 1 and 96 VMScluster nodes. The actual maximum is the highest number of VMScluster nodes possible in the current version of OpenVMS. The default value is the limit defined for the database before it was backed up. You cannot specify the Nodes_Max qualifier if you use the Incremental or Area qualifier. 4 Online Online Noonline Specifies that the restore operation be performed while other users are attached to the database. You can specify the online qualifier only with the Area or Just_Corrupt qualifier. The pages to be restored are locked for exclusive access, so the operation is not compatible with any other use of the data in the specified pages. The default is the Noonline qualifier. 4 Open_Mode Open_Mode=Automatic Open_Mode=Manual Allows you to change the mode for opening a database when you restore that database. When you specify Open_Mode=Automatic, users can invoke the database immediately after it is restored. If you specify Open_Mode=Manual, an RMU Open command must be used to open the database before users can invoke the database. The Open_Mode qualifier also specifies the mode for closing a database. If you specify Open_Mode=Automatic, you can also use the Close_Wait qualifier to specify a time in minutes before the database is automatically closed. If you do not specify the Open_Mode qualifier, the database is restored with the open mode of the database that was in effect when the database was backed up. 4 Options Options=file-spec Specifies the options file that contains storage area names, followed by the storage area qualifiers that you want applied to that storage area. You can direct RMU Restore to create an options file for use with this qualifier by specifying the Restore_Options qualifier with the RMU Backup, RMU Dump, and RMU Dump Backup commands. See Backup Database, Dump Database, and Dump Backup_File for details. If you create your own options file, do not separate the storage area names with commas. Instead, put each storage area name on a separate line in the file. You can include any or all of the area qualifiers in the options file. (See the format help entry under this command for the list of Area qualifiers.) You can use the DCL line continuation character, a hyphen (-), or the comment character (!) in the options file. The default file extension is .opt. 4 Page_Buffers Page_Buffers=number-buffers Specifies the maximum number of buffers Oracle Rdb uses during the RMU Restore operation while the database files are being created. The value of the Page_Buffers qualifier can range from 1 to 5. The default is 3 buffers. Values larger than 3 might improve performance, especially during incremental restore operations. When RMU Restore enters the stage of reconstructing internal structures at the end of the restore operation, a high value for the Page_Buffers qualifier can be useful for very large databases. However, the cost of using these extra buffers is that memory use is high. Thus, the trade-off during a restore operation is memory use against performance. 4 Path Path=cdd-path Specifies a data dictionary path into which the database definitions be integrated. If you do not specify the Path qualifier, RMU Restore uses the CDD$DEFAULT logical name value of the user who entered the RMU Restore command. If you specify a relative path name, Oracle Rdb appends the relative path name you enter to the CDD$DEFAULT value. If the cdd-path parameter contains nonalphanumeric characters, you must enclose it within quotation marks (""). Oracle Rdb ignores the Path qualifier if you use the Nocdd_ Integrate qualifier or if the data dictionary is not installed on your system. 4 Prompt Prompt=Automatic Prompt=Operator Prompt=Client Specifies where server prompts are to be sent. When you specify Prompt=Automatic, prompts are sent to the standard input device, and when you specify Prompt=Operator, prompts are sent to the server console. When you specify Prompt=Client, prompts are sent to the client system. 4 Recovery Recovery[=Aij_Buffers=n] Norecovery The Recovery=Aij_Buffers=n qualifier allows you to specify the number of recovery buffers to use during an automatic recovery. The default value of n is 100 recovery buffers. The Recovery qualifier explicitly specifies that RMU Restore should attempt an automatic recovery of the .aij files during the restore operation. Specify either the Recover=Aij_Buffers=n qualifier and the Recovery qualifier only if .aij files are being retained. If you specify either qualifier in a situation where .aij files are not retained (the Aij_Options, After_Journal, or Duplicate qualifier has been specified), a warning message is displayed and RMU Restore performs the restore operation without attempting to recover the .aij files. The Norecovery qualifier specifies that RMU Restore should not attempt an automatic recovery of the .aij files during the restore operation. Specify this qualifier if you want to use the RMU Recover command with the Until qualifier or if you intend to perform an incremental restore operation. 4 Rewind Rewind Norewind Specifies that the tape that contains the backup file will be rewound before processing begins. The Norewind qualifier, the default, causes the search for the backup file to begin at the current tape position. The Rewind and Norewind qualifiers are applicable only to tape devices. RMU Restore returns an error message if you use these qualifiers and the target device is not a tape device. 4 Root Root=root-file-spec Specifies the database root (.rdb) file specification of the restored database. See the Usage Notes for information on how this qualifier interacts with the Directory, File, and Snapshot qualifiers and for warnings regarding restoring database files into a directory owned by a resource identifier. The Root qualifier is only meaningful when used with a multifile database. 4 Transaction_Mode Transaction_Mode=(mode-list) Sets the allowable transaction modes for the database root file restored by the restore operation. The primary use of this qualifier is when you restore a backup file (of a master database) to create a Hot Standby database. Because only read- only transactions are allowed on a standby database, you should use the Transaction_Mode=Read_Only qualifier setting. This setting prevents modifications to the standby database at all times, even when replication operations are not active. For more information on Hot Standby see the Oracle Rdb7 and Oracle CODASYL DBMS: Guide to Hot Standby Databases. The mode-list can include one or more of the following transaction modes: o All - Enables all transaction modes o Current - Enables all transaction modes that are set for the source database. This is the default transaction mode. o None - Disables all transaction modes o [No]Batch_Update o [No]Read_Only o [No]Exclusive o [No]Exclusive_Read o [No]Exclusive_Write o [No]Protected o [No]Protected_Read o [No]Protected_Write o [No]Read_Write o [No]Shared o [No]Shared_Read o [No]Shared_Write Your restore operation must include the database root file. Otherwise, RMU Restore returns the CONFLSWIT error when you issue an RMU Restore command with the Transaction_Mode qualifier. If you specify more than one transaction mode in the mode-list, enclose the list in parenthesis and separate the transaction modes from one another with a comma. Note the following: o When you specify a negated transaction mode, it indicates that a mode is not an allowable access mode. For example, if you specify the Noexclusive_Write access mode, it indicates that exclusive write is not an allowable access mode for the restored database. o If you specify the Shared, Exclusive, or Protected transaction mode, Oracle RMU assumes you are referring to both reading and writing in that transaction mode. o No mode is enabled unless you add that mode to the list, or you use the All option to enable all transaction modes. o You can list one transaction mode that enables or disables a particular mode followed by another that does the opposite. For example, Transaction_Mode=(Noshared_Write, Shared) is ambiguous because the first value disables Shared_Write access and the second value enables Shared_Write access. Oracle RMU resolves the ambiguity by first enabling the modes as specified in the modes-list and then disabling the modes as specified in the modes-list. The order of items in the list is irrelevant. In the example presented previously, Shared_Read is enabled and Shared_Write is disabled. 4 Users_Max Users_Max=number-users Specifies a new upper limit on the number of users that can simultaneously access the restored database. The valid range is between 1 and 2032 users. The default value is the value defined for the database before it was backed up. You cannot specify the Users_Max qualifier if you use the Incremental qualifier or the Area qualifier. 4 Volumes Volumes = n Allows you to specify that concurrent tape access is to be used to accelerate the restore operation. The Volumes qualifier indicates concurrent tape access and specifies the number of tape volumes in the backup file. The number of volumes must be specified accurately for the restore operation to complete. If you are restoring from a multidisk backup file, the value of "n" indicates the number of disk devices containing backup files needed for the restore operation. If you do not specify the Volumes qualifier, the restore operation does not use concurrent tape access. 4 Blocks_Per_Page Blocks_Per_Page=integer Lets you restore a database with larger mixed page sizes than existed in the original database. This creates new free space on each page in the storage area file and does not interfere with record clustering. RMU Restore ignores this qualifier when it specifies an integer less than or equal to the current page size of the area. You might want to increase the page size in storage areas containing hash indexes that are close to full. By increasing the page size in such a situation, you prevent the storage area from extending. 4 Extension Extension=Disable Extension=Enable Allows you to change the automatic file extension attribute when you restore a database. These qualifiers are positional qualifiers. Use the Extension=Disable qualifier to disable automatic file extension for a storage area. Use the Extension=Enable qualifier to enable automatic file extension for a storage area. If you do not specify the Extension=Disable or Extension=Enable qualifier, the storage areas are restored with the automatic file extension attributes that were in effect when the database was backed up. 4 File File=file-spec Requests that the storage area to which this qualifier is applied be restored in the specified location. This qualifier is not valid for single-file databases. This is a positional qualifier. See the Usage Notes for information on how this qualifier interacts with the Root, Directory, and Snapshot qualifiers and for warnings regarding restoring database files into a directory owned by a resource identifier. 4 Just_Corrupt Just_Corrupt This qualifier replaces the Just_Pages qualifier beginning in Oracle Rdb V7.0. Allows you to restore the corrupt pages and areas in the database as recorded in the corrupt page table (CPT). The CPT is maintained in the .rdb file. (Note that if the corrupt page table becomes full, the area with the highest number of corrupt pages is marked corrupt and the individual pages for that area are removed from the CPT.) Often, only one or a few pages in the database are corrupted due to hardware or software faults. The Just_Corrupt qualifier allows you to recover that database in minimal time with minimal interference; availability of the uncorrupted data is unaffected. It allows you to restrict the restoration to the pages (or areas) logged as corrupt in the corrupt page table. The Just_Corrupt qualifier is a positional qualifier. If you use it in the global position, RMU Restore restores all the corrupt pages and all the corrupt areas as logged in the corrupt page table. If you use it in the local position, RMU Restore restores only the corrupt pages (or the entire area) of the area name it modifies. It is possible to mix restoration of complete areas and just corrupt pages in the same command. The following example restores all of AREA_1 (regardless of whether or not it is corrupt), but just the corrupt pages (logged to the CPT) in AREA_2. $ RMU/RESTORE/AREA backup_file AREA_1, AREA_2/JUST_CORRUPT Note that when the Just_Corrupt qualifier is used globally, all the corrupt pages logged to the CPT for the areas specified are restored. For example, the following command restores all the corrupt pages logged to the CPT for AREA_1 and AREA_2. (However, if one of the areas specified contains no corruptions, an informational message is displayed and that area is not restored.) $ RMU/RESTORE/JUST_CORRUPT backup_file /AREA AREA_1, AREA_2 Restoration of corrupt pages and area can be performed on line. Online operations lock only the corrupt pages or areas for the duration of the restore operation. The remainder of the storage area can be read or updated by an application. When an entire area is restored on line, applications are locked out of the entire area for the duration of the restore operation. There are some restrictions on the use of the Just_Corrupt qualifier: o The backup file must be a full backup file that contains the selected area. o When space area management (SPAM) pages are restored, RMU Restore rebuilds the SPAM page using information from the range of data pages that the SPAM page manages. o Area bit map (ABM) pages can be restored, but their content is not reconstructed. If ABM pages have been corrupted, regenerate them with the RMU Repair command. o A by-page restore operation is like a by-area restore operation in that after-image journal (AIJ) recovery is required to make the restored data consistent with the rest of the database. Once the pages are restored, access to these restored pages is prohibited until they are made consistent. Inconsistent pages are stored in the corrupt page table (CPT) and have their timestamp field flagged by Oracle Rdb. o You can also use the Just_Corrupt qualifier in a restore options file. However, you cannot use any of the following qualifiers with the Just_Corrupt qualifier (neither within an options file nor on the command line): - Blocks_Per_Page - Extension - File - Incremental - Read_Only - Read_Write - Snapshot - Spams - Thresholds You can use the Just_Corrupt qualifier in conjunction with the Journal=file qualifier to greatly speed up processing of a large tape backup file. When you use the Journal qualifier, only those tapes containing corrupt pages, areas, or both, are mounted and processed. 4 Just_Pages Just_Pages[=(p1,p2,...)] This qualifier is replaced with the Just_Corrupt qualifier beginning in Oracle Rdb V7.0. See the description of the Just_ Corrupt qualifier. 4 Read_Only Use the Read_Only qualifier to change a read/write storage area or a write-once storage area to a read-only storage area. If you do not specify the Read_Only or the Read_Write qualifier, the storage areas are restored with the read/write attributes that were in effect when the database was backed up. This is a positional qualifier. 4 Read_Write Use the Read_Write qualifier to change a read-only storage area or a write-once storage area to a read/write storage area. If you do not specify the Read_Only or the Read_Write qualifier, the storage areas are restored with the read/write attributes that were in effect when the database was backed up. This is a positional qualifier. 4 Snapshot Snapshot=(Allocation=n,File=file-spec) If you specify the Allocation parameter, specifies the snapshot file allocation size in n pages for a restored area. If you specify the File parameter, specifies a new snapshot file location for the restored storage area to which it is applied. You can specify the Allocation parameter only, the File parameter only, or both parameters; however, if you specify the Snapshots qualifier, you must specify at least one parameter. This is one of the commands used to alter the parameters of the restored database from those defined at the time of the database backup. Others are /DIRECTORY, /ROOT and /FILE. See the Usage Notes for information on how this qualifier interacts with the Root, File, and Directory qualifiers. The Shapshot qualifier is a positional qualifier. It can be used locally or globally, depending on where the qualifier is placed on the command line. See Examples 22 and 23. To save read/write disk space, you can specify that less space be allocated for the storage area's .snp file when it remains as a read/write file on a read/write disk. If the keyword Allocation is omitted, the original allocation is used. This qualifier is not valid for single-file databases. You cannot specify an .snp file name for a single-file database. When you create an .snp file for a single-file database, Oracle Rdb does not store the file specification of the .snp file. Instead, it uses the file specification of the database root (.rdb) file to determine the file specification of the .snp file. If you want to place the .snp file on a different device or directory, Oracle Corporation recommends that you create a multifile database. However, you can work around the restriction by defining a search list for a concealed logical name. (However, do not use a nonconcealed rooted logical name to define database files; a database created with a non-concealed rooted logical name can be backed up, but may not restore correctly when you attempt to restore the files to a new directory.) To create a database with an .snp file on a different device or directory, define a search list by using a concealed logical name. Specify the location of the root file as the first item in the search list. When you create the database, use the logical name for the directory specification. Then, copy the .snp file to the second device. The following example demonstrates the workaround: $ ! Define a concealed logical name. $ DEFINE /TRANS=CONCEALED/SYSTEM TESTDB USER$DISK1:[DATABASE], - _$ USER$DISK2:[SNAPSHOT] $ $ SQL SQL> -- Create the database. SQL> -- SQL> CREATE DATABASE FILENAME TESTDB:TEST; SQL> EXIT $ ! $ ! Copy the snapshot (.snp) file to the second disk. $ COPY USER$DISK1:[DATABASE]TEST.SNP - _$ USER$DISK2:[SNAPSHOT]TEST.SNP $ ! $ ! Delete the snapshot (.snp) file from the original disk. $ DELETE USER$DISK1:[DATABASE]TEST.SNP; 4 Spams Spams Nospams Enables the space area management (SPAM) pages for the specified area. The Nospams qualifier disables the SPAM pages for the specified area. The default is to leave the attribute unchanged. The Spams and Nospams qualifiers are not allowed for a storage area that has a uniform page format. This is a positional qualifier. 4 Thresholds Thresholds=(val1[,val2[,val3]]) Specifies a storage area's fullness percentage threshold. You can adjust SPAM thresholds to improve future space utilization in the storage area. Each threshold value represents a percentage of fullness on a data page. When a data page reaches the percentage of fullness defined by a given threshold value, the space management entry for the data page is updated to contain that threshold value. The Thresholds qualifier applies only to storage areas with a mixed page format. If you do not use the Thresholds qualifier with the RMU Restore command, Oracle Rdb uses the storage area's original thresholds. This is a positional qualifier. See the Oracle Rdb7 Guide to Database Performance and Tuning for more information on setting SPAM thresholds. 3 Usage_Notes o To use the RMU Restore command for a database, you must have the RMU$RESTORE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o The RMU Restore command provides four qualifiers, Directory, Root, File, and Snapshots, that allow you to specify the target for the restored files. The target can be just a directory, just a file name, or a directory and file name. If you use all or some of these four qualifiers, apply them as follows: - Use the Root qualifier to indicate the target for the restored database root file. - Use local application of the File qualifier to specify the target for the restored storage area or areas. - Use local application of the Snapshots qualifier to specify the target for the restored snapshot file or files. - Use the Directory qualifier to specify a default target directory. The default target directory is the directory to which all files not qualified with the Root, File, or Snapshot qualifier are restored. It is also the default directory for files qualified with the Root, File, or Snapshot qualifier if the target for these qualifiers does not include a directory specification. Note the following when using these qualifiers: - Global application of the File qualifier when the target specification includes a file name causes RMU Restore to restore all of the storage areas to different versions of the same file name. This creates a database that is difficult to manage. - Global application of the Snapshot qualifier when the target specification includes a file name causes RMU Restore to restore all of the snapshot files to different versions of the same file name. This creates a database that is difficult to manage. - Specifying a file name or extension with the Directory qualifier is permitted, but causes RMU Restore to restore all of the files (except those specified with the File or Root qualifier) to different versions of the same file name. Again, this creates a database that is difficult to manage. See Example 17. o When you restore a database into a directory owned by a resource identifier, the ACE for the directory is applied to the database root file ACL first, and then the Oracle RMU ACE is added. This method is employed to prevent database users from overriding OpenVMS file security. However, this can result in a database which you consider yours, but to which you have no Oracle RMU privileges to access. See the Oracle Rdb Guide to Database Maintenance for details. o If a backup file to tape is created using a single tape device, it must be restored using a single tape device; it cannot be restored using multiple tape devices. NOTE An incremental backup file created for a database running under one version of Oracle Rdb cannot be applied if that database has been restored under another version of Oracle Rdb. For example, if you do the following, step 6 fails with the error message, "XVERREST, Cross version RESTORE is not possible for by-area or incremental functions": 1. Apply a full backup operation to a Version 7.1 database. 2. Apply updates to the database. 3. Perform an incremental backup operation on the database. 4. Move backup files to a system running Oracle Rdb Version 7.2. 5. Restore the database by using the full backup file. 6. Attempt to apply the incremental backup file created in step 1. o If you apply an incremental backup file, you must specify the Norecovery qualifier when you issue a full RMU Restore command for the corresponding full backup file. o If you mistakenly attempt to restore a backup file in a version of Oracle Rdb that is earlier than the version for which the backup file was created, you might receive INVRECTYP errors and your operation will probably terminate with an access violation (ACCVIO) exception. If you receive this error, check the version of the backup file and the version of Oracle Rdb you are running. Be sure the environment version matches, or is greater than, the version under which the backup file was created. o RMU Restore might create an .rdb file and .rda files when it starts up. If you specify the Log qualifier, these files will be noted in the log file. These are not database files until the end of the operation when they have been populated with the backed-up contents. Therefore, if the restore operation aborts or is stopped using Ctrl/Y, you must delete these unpopulated files by using the DCL DELETE command. You know which files to delete by the contents of the backup file and the form of the command issued, or by examining the output in the log file if you specified the Log qualifier. Deleting the files usually requires OpenVMS privileges. Until they are restored, these files are not a database, and Oracle RMU or SQL operations do not function with them. o RMU Restore preserves any area reservations and after-image journal (.aij) file reservations that exist in the backed-up database. o If you restore a database without its root file ACL (using the Noacl qualifier with the RMU Restore command, for example), a user who wants to create ACL entries for the database must have the OpenVMS SECURITY or BYPASS privilege. o The RMU Restore command with the Area and Online qualifiers requires exclusive access to the area files being restored. The RMU Restore command with the Area, Online, and Just_ Corrupt qualifiers requires exclusive access to only the pages being restored. o There are no restrictions on the use of the Nospams qualifier with storage areas that have a mixed page format, but the use of the Nospams qualifier typically causes severe performance degradation. The Nospams qualifier is useful only where updates are rare and batched, and access is primarily by database key (dbkey). o The RMU Restore command automatically uses the RMU Convert command when restoring the database to a system with a more recent version of Oracle Rdb software. When this is done, the metadata in the Oracle Rdb database changes and invalidates incremental backup files from the previous version. By default, no areas are reserved and one .aij file is reserved. (You can override the after-image journal default reservation by using the Aij_Options qualifier.) See Convert for information on the versions of Oracle Rdb that the Convert command supports. o Always back up your Oracle Rdb databases as recommended in the Oracle Rdb Installation and Configuration Guide just prior to installing a newer version of Oracle Rdb software. The last backup file made prior to converting to a more recent version of Oracle Rdb should be a full and complete backup file. o See the Oracle Rdb Guide to Database Maintenance for information on the steps RMU Restore follows in tape label checking when you restore a database from tape. o RMU Restore might initialize the SPAM thresholds for some data pages of some storage areas that have a uniform page format to values that are not acceptable to the RMU Verify command. This occurs when some of the data pages in a logical area are restored before the logical area definition (Area Inventory). This is not a frequent occurrence, and when it does happen, the consequences are usually cosmetic (the RMU Verify command issues a warning message for each page affected). However, if many pages are affected, the volume of warnings can cause you to overlook a real problem. Moreover, in some cases, this can result in additional I/O operations when new data is stored in an affected table. As a workaround, you can use the RMU Repair command to reconstruct the SPAM pages in one or more storage areas. The RMU Repair command corrects the condition caused by the RMU Restore command as well as other SPAM page corruptions. See the help entry for the RMU Repair command for more information on the RMU Repair command. 3 Examples Example 1 The following example restores the mf_personnel database from the backup file pers_bu.rbf and requests a new version of the database file. Because the After_Journal qualifier has been specified, automatic recovery will not be attempted. $ RMU/RESTORE/NEW_VERSION/AFTER_JOURNAL=AIJ_DISK:[AIJS]PERSAIJ - _$ /NOCDD_INTEGRATE/LOG PERS_BU - _$ EMP_INFO /THRESHOLDS=(65,75,80)/BLOCKS_PER_PAGE=3 The command changes the .aij file location and name to AIJ_DISK:[AIJS]PERSAIJ.AIJ, prevents integration with the data dictionary, and displays the progress of the restore operation. For the storage area, EMP_INFO, the command changes the SPAM threshold values to 65%, 75%, and 80%, and increases the number of blocks per page to 3 blocks. Example 2 Assume that at 10 A.M., Wednesday, October 25, 2005, a disk device hardware failure corrupted all the files on the device, including the mf_personnel.rdb file. The following command restores the full database backup file (pers_full_oct22.rbf) created on the previous Sunday and then restores the incremental backup file made on Tuesday. Note that an incremental database backup file was created on Monday, but each new incremental backup file made since the latest full backup file replaces previous incremental backup files made since the last full backup operation. $ RMU/RESTORE/LOG/NORECOVERY MUA1:PERS_FULL_OCT22.RBF $ RMU/RESTORE/INCREMENTAL/CONFIRM/LOG/NORECOVERY - _$ PERS_INCR_OCT24.RBF At this point, the database is current up until 11:30 P.M., Tuesday, when the last incremental backup file was made of mf_ personnel. Because after-image journaling is enabled for this database, automatic recovery of the .aij file could have been employed. However, if the recovery process should fail for some reason or, as in this case, the Norecovery qualifier is specified, you can still use the RMU Recover command to apply the .aij file that contains changes made to the database from 11:30 P.M., Tuesday, until just before the hardware failure to the restored mf_personnel.rdb file and its storage area files. For example: $ RMU/RECOVER/UNTIL = "25-OCT-2005 09:55:00.00" - _$ AIJ_DISK:[AIJS]PERSAIJ.AIJ;1 Example 3 If a storage area is on a disk that fails, you might want to move that storage area to another disk by using the RMU Restore command. The following RMU Restore command restores only the EMPIDS_OVER storage area from the full backup file of mf_ personnel, and moves the EMPIDS_OVER storage area and snapshot (.snp) file to a new location on the 333$DUA11 disk. The recovery operation is only required if the required .aij file has been backed up and is no longer in the current aij state. $ RMU/RESTORE/AREA 222$DUA20:[BACKUPS]MF_PERS_BU.RBF - _$ EMPIDS_OVER /FILE=333$DUA11:[DBS]EMPIDS_OVER.RDA - _$ /SNAPSHOT=(FILE=333$DUA11:[DBS]EMPIDS_OVER.SNP) $ ! $ ! Recovery from the after-image journal is automatic. If $ ! automatic recovery is not possible, or if the Norecovery $ ! qualifier had been specified, perform the following: $ ! $ RMU/RECOVER/AREA AIJ_DISK:PERS.AIJ Example 4 The following example demonstrates how you can use by-area backup and restore operations for a single storage area in the mf_ personnel database. In addition, it demonstrates the use of the automatic recovery feature of the RMU Restore command. $ ! $ ! Create an .aij file for the database. Because three $ ! .aij files are created, fixed-size .aij $ ! journaling will be used. $ ! $ RMU/SET AFTER_JOURNAL/ENABLE/RESERVE=4 - _$ /ADD=(name=AIJ1, FILE=DISK2:[CORP]AIJ_ONE) - _$ /ADD=(name=AIJ2, FILE=DISK2:[CORP]AIJ_TWO) - _$ /ADD=(NAME=AIJ3, FILE=DISK2:[CORP]AIJ_THREE) - _$ MF_PERSONNEL.RDB %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery $ RMU/BACKUP MF_PERSONNEL DISK3:[BACKUP]MF_PERS.RBF $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL'; SQL> -- SQL> -- On Monday, define a new row in the DEPARTMENTS table. The SQL> -- new row is stored in the DEPARTMENTS storage area. SQL> -- SQL> INSERT INTO DEPARTMENTS cont> (DEPARTMENT_CODE, DEPARTMENT_NAME, MANAGER_ID, cont> BUDGET_PROJECTED, BUDGET_ACTUAL) cont> VALUES ('WLNS', 'Wellness Center', '00188', 0, 0); 1 row inserted SQL> SQL> COMMIT; SQL> EXIT; $ ! $ ! Assume that you know that the only storage area ever updated in $ ! the mf_personnel database on Tuesdays is the SALARY_HISTORY $ ! storage area, and you decide that you will create an incremental $ ! backup file of just the SALARY_HISTORY storage area on Tuesday. $ ! Before you perform the by-area backup operation on the $ ! SALARY_HISTORY storage area on Tuesday, you must perform a full $ ! and complete backup operation on the mf_personnel database when $ ! it is in a known and working state. $ ! $ RMU/BACKUP MF_PERSONNEL.RDB - _$ DISK3:[BACKUP]MF_MONDAY_FULL.RBF $ ! SQL> -- SQL> -- On Tuesday, two rows are updated in SQL> -- the SALARY_HISTORY storage area. SQL> -- SQL> UPDATE SALARY_HISTORY cont> SET SALARY_END ='20-JUL-2003 00:00:00.00' cont> WHERE SALARY_START='14-JAN-1993 00:00:00.00' cont> AND EMPLOYEE_ID = '00164'; 1 row updated SQL> UPDATE SALARY_HISTORY cont> SET SALARY_START ='5-JUL-2000 00:00:00.00' cont> WHERE SALARY_START='5-JUL-1990 00:00:00.00' cont> AND EMPLOYEE_ID = '00164'; 1 row updated SQL> COMMIT; SQL> EXIT; $ ! $ ! On Tuesday, you create an incremental backup file of the $ ! SALARY_HISTORY storage area only. Only the SALARY_HISTORY $ ! storage area is included in the by-area backup file. $ ! Oracle RMU provides an informational message telling $ ! you that not all storage areas in the database are included $ ! in the mf_tuesday_partial.rbf backup file. $ RMU/BACKUP/INCLUDE=(SALARY_HISTORY) - _$ /INCREMENTAL/LOG DISK1:[USER]MF_PERSONNEL.RDB - _$ DISK3:[BACKUPS]MF_TUESDAY_PARTIAL.RBF %RMU-I-NOTALLARE, Not all areas will be included in this backup file %RMU-I-LOGLASCOM, Last full and complete backup was dated 18-JAN-2006 11:19:46.31 %RMU-I-BCKTXT_00, Backed up root file DISK1:[DB]MF_PERSONNEL.RDB;1 %RMU-I-BCKTXT_03, Starting incremental backup of storage area DISK3:[SA}SALARY_HISTORY.RDA;1 at 18-JAN-2006 11:20:49.29 %RMU-I-BCKTXT_13, Completed incremental backup of storage area DISK3:[SA]SALARY_HISTORY.RDA;1 at 18-JAN-2006 11:20:49.40 %RMU-I-COMPLETED, BACKUP operation completed at 18-JSN-2006 11:20:49.59 . . . $ ! SQL> -- Update another row in the SALARY_HISTORY table: SQL> UPDATE SALARY_HISTORY cont> SET SALARY_START ='23-SEP-1991 00:00:00.00' cont> WHERE SALARY_START='21-SEP-1981 00:00:00.00' cont> AND EMPLOYEE_ID = '00164'; 1 row updated SQL> COMMIT; SQL> EXIT; $ ! Assume that a disk device hardware error occurs here $ ! and only the SALARY_HISTORY storage area and snapshot $ ! file is lost. Also assume that the database root (.rdb) $ ! file and other storage areas in the database are still $ ! fine and do not need to be restored or recovered. $ ! Therefore, you do not need to restore the .rdb file or $ ! other storage areas from the full and complete backup $ ! file. Because only the SALARY_HISTORY storage area was $ ! lost, you must do the following: $ ! 1) Restore the SALARY_HISTORY storage area and snapshot $ ! file from the last full and complete backup file. Note $ ! this operation can be done on line. Specify the Norecovery $ ! qualifier because you still have an incremental restore $ ! operation to perform. $ ! 2) Restore the SALARY_HISTORY storage area from the last $ ! incremental backup file. Note this operation can be $ ! done on line. This time do not specify the Norecovery $ ! qualifier so that the automatic recovery provided by $ ! Oracle RMU will be implemented. $ ! $ RMU/RESTORE/NOCDD_INTEGRATE/ONLINE/LOG/NORECOVERY - _$ /AREA DISK3:[BACKUP]MF_MONDAY_FULL.RBF SALARY_HISTORY %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]SALARY_HISTORY.RDA;1 at 18-JAN-2006 11:25:13.17 %RMU-I-RESTXT_24, Completed full restore of storage area DISK1:[USER]SALARY_HISTORY.RDA;1 at 18-JAN-2006 11:25:13.86 %RMU-I-RESTXT_01, Initialized snapshot file DISK1:[USER]SALARY_HISTORY.SNP;1 %RMU-I-LOGINIFIL, contains 100 pages, each page is 2 blocks long %RMU-I-AIJWASON, AIJ journaling was active when the database was backed up %RMU-I-AIJRECFUL, Recovery of the entire database starts with AIJ file sequence 0 %RMU-I-AIJRECARE, Recovery of area SALARY_HISTORY starts with AIJ file sequence 0 %RMU-I-COMPLETED, RESTORE operation completed at 18-JAN-2006 11:25:14.51 $ RMU/RESTORE/NOCDD_INTEGRATE/INCREMENTAL/ONLINE/LOG - _$ /AREA DISK3:[BACKUPS]MF_TUESDAY_PARTIAL.RBF SALARY_HISTORY DISK1:[USER]MF_PERSONNEL.RDB;1, restore incrementally? [N]:Y %RMU-I-RESTXT_22, Starting incremental restore of storage area DISK1:[USER]SALARY_HISTORY.RDA;1 at 18-JAN-2006 11:29:35.54 %RMU-I-RESTXT_25, Completed incremental restore of storage area DISK1:[USER]SALARY_HISTORY.RDA;1 at 18-JAN-2006 11:29:35.64 %RMU-I-RESTXT_01, Initialized snapshot file DISK1:[USER]SALARY_HISTORY.SNP;1 %RMU-I-LOGINIFIL, contains 100 pages, each page is 2 blocks long %RMU-I-AIJWASON, AIJ journaling was active when the database was backed up %RMU-I-AIJRECFUL, Recovery of the entire database starts with AIJ file sequence 0 %RMU-I-AIJRECARE, Recovery of area SALARY_HISTORY starts with AIJ file sequence 0 %RMU-I-AIJBADAREA, inconsistent storage area DISK1:[USER]SALARY_HISTORY.RDA;1 needs AIJ sequence number 0 %RMU-I-LOGRECDB, recovering database file DISK1:[USER]MF_PERSONNEL.RDB;1 %RMU-I-AIJAUTOREC, starting automatic after-image journal recovery %RMU-I-LOGOPNAIJ, opened journal file DISK2:[CORP]AIJ_ONE.AIJ;17 %RMU-I-AIJONEDONE, AIJ file sequence 0 roll-forward operations completed %RMU-I-LOGRECOVR, 1 transaction committed %RMU-I-LOGRECOVR, 0 transactions rolled back %RMU-I-LOGRECOVR, 3 transactions ignored %RMU-I-AIJNOACTIVE, there are no active transactions %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-LOGSUMMARY, total 1 transaction committed %RMU-I-LOGSUMMARY, total 0 transactions rolled back %RMU-I-LOGSUMMARY, total 3 transactions ignored %RMU-I-AIJSUCCES, database recovery completed successfully Example 5 In the following example, the options file specifies that the storage area (.rda) files are to be restored to different disks. Note that storage area snapshot (.snp) files are restored to different disks from one another and from their associated storage area (.rda) files; this is recommended for optimal performance. (This example assumes that the disks specified for each storage area file in options_file.opt are different from those where the storage area files currently reside.) $ RMU/RESTORE/NOCDD_INTEGRATE/OPTIONS=OPTIONS_FILE.OPT - _$ MF_PERS_BCK.RBF $ TYPE OPTIONS_FILE.OPT EMPIDS_LOW /FILE=DISK1:[CORPORATE.PERSONNEL]EMPIDS_LOW.RDA - /SNAPSHOT=(FILE=DISK2:[CORPORATE.PERSONNEL]EMPIDS_LOW.SNP ) EMPIDS_MID /FILE=DISK3:[CORPORATE.PERSONNEL]EMPIDS_MID.RDA - /SNAPSHOT=(FILE=DISK4:[CORPORATE.PERSONNEL]EMPIDS_MID.SNP ) EMPIDS_OVER /FILE=DISK5:[CORPORATE.PERSONNEL]EMPIDS_OVER.RDA - /SNAPSHOT=(FILE=DISK6:[CORPORATE.PERSONNEL]EMPIDS_OVER.SNP ) DEPARTMENTS /FILE=DISK7:[CORPORATE.PERSONNEL]DEPARTMENTS.RDA - /SNAPSHOT=(FILE=DISK8:[CORPORATE.PERSONNEL]DEPARTMENTS.SNP ) SALARY_HISTORY /FILE=DISK9:[CORPORATE.PERSONNEL]SALARY_HISTORY.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]SALARY_HISTORY.SNP ) JOBS /FILE=DISK7:[CORPORATE.PERSONNEL]JOBS.RDA - /SNAPSHOT=(FILE=DISK8:[CORPORATE.PERSONNEL]JOBS.SNP ) EMP_INFO /FILE=DISK9:[CORPORATE.PERSONNEL]EMP_INFO.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]EMP_INFO.SNP ) RESUME_LISTS /FILE=DISK11:[CORPORATE.PERSONNEL]RESUME_LISTS.RDA - /SNAPSHOT=(FILE=DISK12:[CORPORATE.PERSONNEL]RESUME_LISTS.SNP ) RESUMES /FILE=DISK9:[CORPORATE.PERSONNEL]RESUMES.RDA - /SNAPSHOT=(FILE=DISK10:[CORPORATE.PERSONNEL]RESUMES.SNP ) Example 6 The following example shows what .aij file sequence to use following an RMU Restore command with the Area qualifier if automatic recovery fails: $ RMU/RESTORE/AREA MFPERS_62691.RBF - DEPARTMENTS, JOBS . . . %RMU-I-AIJWASON, AIJ journaling was active when the database was backed up %RMU-I-AIJRECFUL, Recovery of the entire database starts with AIJ file sequence 0 Example 7 The following example shows how to move a single-file database to a new directory, using the RMU Backup and RMU Restore commands: $ RMU/BACKUP PERSONNEL PERS $! $ RMU/RESTORE/NOCDD/NOAFTER_JOURNAL - _$ /DIRECTORY=DISK4:[USER2] PERS Example 8 The following example shows how to rename a single-file database when you move the database by using the RMU Backup and RMU Restore commands: $ RMU/BACKUP PERSONNEL PERS $! $ RMU/RESTORE/NOCDD/NOAFTER_JOURNAL - _$ /DIRECTORY=DISK4:[USER2]TEST_PERSONNEL PERS Example 9 The following example causes the database being restored from the mf_pers_bck.rbf backup file to have 60 global buffers, with a limit of 2 buffers for each database user. Because the Enabled option is used, global buffering is in effect for the database immediately after it is restored: $ RMU/RESTORE/NOCDD/GLOBAL_BUFFERS=(ENABLED,TOTAL=60,USER_LIMIT=2) - _$ MF_PERS_BCK.RBF Example 10 The following command causes the SALARY_HISTORY storage area from the database being restored from the mf_pers_bu.rbf backup file to be restored as a read-only storage area. None of the other database storage areas are modified as part of this restore operation. $ RMU/RESTORE/NOCDD MF_PERS_BU.RBF SALARY_HISTORY /READ_ONLY Example 11 The following example assumes that you are using multiple tape drives to perform a large restore operation. By specifying the Loader_Synchronization and Volumes qualifiers, this command does not require you to load tapes as each completes. Instead, you can load tapes on a loader or stacker and the RMU restore process will wait until all concurrent tape operations have concluded for one set of tape volumes before assigning the next set of tape volumes. This example assumes that the backup operation used two tape output threads and each thread wrote four tapes. This example uses Master qualifiers to indicate that you want the $111$MUA0: and $444$MUA2: drives to be master drives. Using this example, you would: 1. Allocate each tape drive. 2. Manually place tapes BACK01 and BACK05 on the $111$MUA0: master drive. 3. Manually place tapes BACK02 and BACK06 on the $333$MUA2: master drive. 4. Manually place tapes BACK03 and BACK07 on the $222$MUA1: slave drive. 5. Manually place tapes BACK04 and BACK08 on the $444$MUA3: slave drive. 6. Mount the first volume (BACK01). 7. Perform the restore operation. 8. Dismount the last tape mounted. 9. Deallocate each tape drive. $ ALLOCATE $111$MUA0: $ ALLOCATE $222$MUA1: $ ALLOCATE $333$MUA2: $ ALLOCATE $444$MUA3: $ $ MOUNT/FOREIGN $111$MUA0: $ $ RMU/RESTORE/LOG/REWIND/LOADER_SYNCHRONIZATION - _$ /LABEL=(BACK01, BACK02, BACK03, BACK04, BACK05, - _$ BACK06, BACK07, BACK08) - _$ /VOLUMES=8 - _$ $111$MUA0:PERS_FULL_MAR30.RBF/MASTER, $222$MUA1: - _$ $333$MUA2:/MASTER, $444$MUA3 $ $ DISMOUNT $222$MUA3: $ $ DEALLOCATE $111$MUA0: $ DEALLOCATE $222$MUA1: $ DEALLOCATE $333$MUA2: $ DEALLOCATE $444$MUA3: Example 12 The following example demonstrates the automatic .aij recovery mechanism in the RMU Restore command. The example does the following: o Uses the RMU Set After_Journal command to reserve space for four .aij files, adds three .aij files, and enables after- image journaling o Performs a backup operation on the database o Performs database update activity, which will be written to an .aij file o Determines the database root file is lost o Restores and recovers the database in one RMU Restore command $ SET DEFAULT DISK1:[USER] $ ! $ RMU/SET AFTER_JOURNAL/ENABLE/RESERVE=4 - _$ /ADD=(name=AIJ1, FILE=DISK2:[CORP]AIJ_ONE) - _$ /ADD=(name=AIJ2, FILE=DISK2:[CORP]AIJ_TWO) - _$ /ADD=(NAME=AIJ3, FILE=DISK2:[CORP]AIJ_THREE) - _$ MF_PERSONNEL %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery $ ! $ ! Back up database, as instructed. $ ! $ RMU/BACKUP MF_PERSONNEL DISK3:[BACKUPS]MF_PERS.RBF $ ! $ ! Database update activity occurs. $ ! $! $! Database is lost. Issue the RMU Restore command to $! restore and recover the database. Because the Norecovery $! qualifier is not specified, Oracle RMU will $! automatically attempt to recover the database. $! $ RMU/RESTORE DISK3:[BACKUPS]MF_PERS.RBF/NOCDD_INTEGRATE %RMU-I-AIJRSTAVL, 3 after-image journals available for use %RMU-I-AIJRSTMOD, 1 after-image journal marked as "modified" %RMU-I-AIJISON, after-image journaling has been enabled %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery %RMU-I-LOGRECDB, recovering database file DISK1:[USER]MF_PERSONNEL.RDB;1 %RMU-I-AIJAUTOREC, starting automatic after-image journal recovery %RMU-I-AIJONEDONE, AIJ file sequence 0 roll-forward operations completed %RMU-I-AIJONEDONE, AIJ file sequence 1 roll-forward operations completed %RMU-W-NOTRANAPP, no transactions in this journal were applied %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJFNLSEQ, to start another AIJ file recovery, the sequence number needed will be 1 Example 13 The following example demonstrates how to restore and recover all the corrupt pages and areas in the mf_personnel database. Assume that the RMU Show Corrupt_Pages command shows that the JOBS storage area is corrupt and that only page 3 in the DEPARTMENTS storage area is corrupt. All the other storage areas are neither corrupt nor inconsistent. Because the Just_Corrupt qualifier is specified in the global position, and mf_personnel.rbf is a full backup file, the RMU restore process restores all of the JOBS storage area and just page 3 in the DEPARTMENTS storage area. If after-image journaling is enabled, automatic recovery will be attempted. $ RMU/RESTORE/AREA/JUST_CORRUPT MF_PERSONNEL.RBF Example 14 The following example demonstrates how to restore and recover specific corruptions in the mf_personnel database. Like example 12, assume that the RMU Show Corrupt_Pages command shows that the JOBS storage area is corrupt and that only page 3 in the DEPARTMENTS storage area is corrupt. All the other storage areas are neither corrupt nor inconsistent. The backup file, mf_partial.rbf, is a by-area backup file containing backups of the JOBS, DEPARTMENTS, and SALARY_HISTORY storage areas. In this example, the JOBS, DEPARTMENTS, and SALARY_HISTORY areas are specified for restoring. Because the SALARY_HISTORY area contains no corruptions, an informational message is returned. The RMU restore process restores all of the JOBS storage area and just page 3 in the DEPARTMENTS storage area. If after-image journaling is enabled, automatic recovery will be attempted. $ RMU/RESTORE/JUST_CORRUPT/AREA MF_PARTIAL.RBF JOBS, - _$ DEPARTMENTS,SALARY_HISTORY %RMU-I-RESTXT_20, Storage area DISK1:[AREA]SALARY_HISTORY.RDA;1 is not corrupt and will not be restored Example 15 The following example demonstrates how to restore and recover specific corruptions in the mf_personnel database along with restoring an area that is not corrupt. Like example 13, assume that the RMU Show Corrupt_Pages command shows that the JOBS storage area is corrupt and that only page 3 in the DEPARTMENTS storage area is corrupt. All the other storage areas are neither corrupt nor inconsistent. The backup file, mf_personnel.rbf, is a full backup file. In this example, the Just_Corrupt qualifier is used locally with the DEPARTMENTS storage area. The JOBS, DEPARTMENTS, and SALARY_HISTORY areas are specified for restoring. Although the SALARY_HISTORY area contains no corruptions, an informational message is not returned in this case because by specifying the Just_Corrupt qualifier locally with DEPARTMENTS, the Restore command is requesting that the RMU restore process restore the JOBS and SALARY_HISTORY storage areas regardless of corruptions, and the DEPARTMENTS storage area be restored to fix corruptions. The RMU restore process restores all of the JOBS and SALARY_HISTORY storage areas and just page 3 in the DEPARTMENTS storage area. If after-image journaling is enabled, automatic recovery will be attempted. $ RMU/RESTORE/AREA MF_PERSONNEL.RBF JOBS, SALARY_HISTORY, - _$ DEPARTMENTS/JUST_CORRUPT Example 16 The following example is the same as example 15, except the Just_ Corrupt qualifier is specified locally with the SALARY_HISTORY storage area. Because the SALARY_HISTORY qualifier contains no corruptions, an error message is returned: $ RMU/RESTORE/AREA MF_PERSONNEL.RBF JOBS,SALARY_HISTORY/JUST_CORRUPT, - _$ DEPARTMENTS/JUST_CORRUPT %RMU-I-RESTXT_20, Storage area DISK1:[AREA]SALARY_HISTORY.RDA;1 is not corrupt and will not be restored Example 17 The following example demonstrates the behavior of the RMU Restore command when the Just_Corrupt qualifier is used both globally and locally. The global use of the Just_Corrupt qualifier overrides an local use of the qualifier. In this case, the RMU restore process restores the JOBS, SALARY_HISTORY, and DEPARTMENTS storage areas only if they contain corruptions; otherwise an error is returned. Assume, like the previous examples, that only the JOBS and DEPARTMENTS storage areas contain corruptions: $ RMU/RESTORE/JUST_CORRUPT/AREA MF_PERSONNEL.RBF SALARY_HISTORY, - _$ JOBS/JUST_CORRUPT, DEPARTMENTS/JUST_CORRUPT %RMU-I-RESTXT_20, Storage area DISK1:[AREA]SALARY_HISTORY.RDA;1 is not corrupt and will not be restored 3 Examples_(Cont.) Example 18 The following example demonstrates the use of the Directory, File, and Root qualifiers. In this example: o The default directory is specified as DISK2:[DIR]. o The target directory and file name for the database root file is specified with the Root qualifier. The target directory specified with the Root qualifier overrides the default directory specified with the Directory qualifier. Thus, the RMU restore process restores the database root in DISK3:[ROOT] and names it COPYRDB.RDB. o The target directory for the EMPIDS_MID storage area is DISK4:[FILE]. The RMU restore process restores EMPIDS_MID in DISK4:[FILE]. o The target file name for the EMPIDS_LOW storage area is EMPIDS. Thus, the RMU restore process restores the EMPIDS_LOW storage area to the DISK2:[DIR] default directory (specified with the Directory qualifier), and names the file EMPIDS.RDA. o The target for the EMPIDS_LOW snapshot file is DISK5:[SNAP]EMPIDS.SNP. Thus, the RMU restore process restores the EMPIDS_LOW snapshot file to DISK5:[SNAP]EMPIDS.SNP. o All the other storage area files and snapshot files in the mf_ personnel database are restored in DISK2:[DIR]; the file names for these storage areas and snapshot files remain unchanged. $ RMU/RESTORE MF_PERSONNEL.RBF - _$ /DIRECTORY=DISK2:[DIR] - _$ /ROOT=DISK3:[ROOT]MF_PERSONNEL.RDB - _$ EMPIDS_MID/FILE=DISK4:[FILE], - _$ EMPIDS_LOW/FILE=EMPIDS - _$ /SNAPSHOT=(FILE=DISK5:[SNAP]EMPIDS.SNP) Example 19 The following example demonstrates how to restore a database such that the newly restored database will allow read-only transactions only. After the RMU restore process executes the command, the database is ready for you to start Hot Standby replication operations. See the Oracle Rdb7 and Oracle CODASYL DBMS: Guide to Hot Standby Databases for details on starting Hot Standby replication operations. $RMU/RESTORE/TRANSACTION_MODE=READ_ONLY MF_PERSONNEL.RBF Example 20 The following example uses the Nocommit qualifier while restoring a backup file of a database that has a structure level of V7.1 in a V7.2 environment. $ RMU/SHOW VERSION Executing RMU for Oracle Rdb V7.2-00 $ RMU/RESTORE MFP71.RBF /NOCOMMIT/NOCDD/NORECOVER %RMU-I-AIJRSTAVL, 0 after-image journals available for use %RMU-I-AIJISOFF, after-image journaling has been disabled %RMU-I-LOGCONVRT, database root converted to current structure level %RMU-S-CVTDBSUC, database USER1:[80]MF_PERSONNEL.RDB;1 successfully converted from version V7.1 to V7.2 %RMU-W-USERECCOM, Use the RMU Recover command. The journals are not available. $ RMU/SHOW VERSION Executing RMU for Oracle Rdb V7.2-00 $ RMU/CONVERT/ROLLBACK MF_PERSONNEL.RDB %RMU-I-RMUTXT_000, Executing RMU for Oracle Rdb V7.2-00 Are you satisfied with your backup of RDBVMS_USER1:[V71]MF_PERSONNEL.RDB;1 and your backup of any associated .aij files [N]? Y %RMU-I-LOGCONVRT, database root converted to current structure level %RMU-I-CVTROLSUC, CONVERT rolled-back for RDBVMS_USER1:[V71]MF_PERSONNEL. RDB;1 to version V7.1 Example 21 The following example uses the Close_Wait qualifier to set the database close mode to TIMED AUTOMATIC, specifying that the database will be closed automatically in 10 minutes. $ RMU/RESTORE/OPEN_MODE=AUTOMATIC/CLOSE_WAIT=10/DIR=DISK:[DIR] TEST_DB.RBF $ RMU/DUMP/HEADER=PARAMETERS TEST_DB.RDB Example 22 The following example demonstrates that /SNAPSHOT=(ALLOCATION=N) is a positional qualifier. The behavior is different (local or global) depending on the placement of the qualifier on the command line. In the following example, it is used both globally and locally. MALIBU-> RMU/RESTORE/NOCDD - /DIR=SYS$DISK:[]/SNAP=ALLO=12345 [JONES.RDB]MF_PERSONNEL_V71.RDF - DEPARTMENTS/SNAP=ALLO=2 MALIBU-> DIR/SIZE *.SNP Directory DBMS_USER3:[JONES.WORK] DEPARTMENTS.SNP;1 6 EMPIDS_LOW.SNP;1 24692 EMPIDS_MID.SNP;1 24692 EMPIDS_OVER.SNP;1 24692 EMP_INFO.SNP;1 24692 JOBS.SNP;1 24692 MF_PERS_DEFAULT.SNP;1 24692 MF_PERS_SEGSTR.SNP;1 24692 SALARY_HISTORY.SNP;1 24692 Total of 9 files, 197542 blocks. Example 23 The following example demonstrates how /SNAPSHOT=(ALLOCATION=N) can be used to alter the parameters of the restored database from those defined at the time of the database backup. /SNAPSHOT is ofter used with /FILE: /FILE for the storage area RDA file and /SNAPSHOT for the storage area snapshot file. $ RMU/RESTORE MFP.RBF - /DIRECTORY=DISK1:[DIRECTORY] - /ROOT=DISK2:[DIRECTORY]MF_PERSONNEL.RDB - EMPIDS_MID /FILE=[DISK3:[DIRECTORY] /SNAPSHOT=(ALLOCATION=2000), - EMPIDS_LOW /FILE=[DISK3:[DIRECTORY]NEWNAME - /SNAPSHOT=(FILE=DISK4:[DIR]NEWNAME, ALLOCATION=3000) In this example, the root would go to one disk, EMPIDS_MID would go to another, EMPIDS_LOW to another disk and the snap to another disk and both snaps would be allocated the specified number of pages. All the other snaps and RDA files would go to where /DIRECTORY points (and the snaps would keep their original allocation). 3 Only_Root Permits you to recover more quickly from the loss of a database root (.rdb) file by restoring only the root file. This command is not valid for single-file databases. 4 Description The RMU Restore Only_Root command rebuilds only the database root (.rdb) file from a backup file, produced earlier by an RMU Backup command, to the condition the .rdb file was in when the backup operation was performed. Use the command qualifiers to update the .rdb file. The area qualifiers alter only the .rdb file, not the storage areas themselves. Use the area qualifiers to correct the restored backup root file so that it contains storage area information that was updated since the last backup operation was performed on the database. This is useful when you need to match the root from an older backup file of your database with the area information in the more recent backup file of your database in order to have a usable database. When the .rdb file is restored by itself, be sure that you correctly set the transaction state of the database with the Initialize_Tsns qualifier or the Set_Tsn qualifier. If the database transaction sequence number (TSN) and commit sequence number (CSN) are not set to the same values as those that were in the lost .rdb file, there will be an inconsistency in the journaling if after-image journaling is enabled. Therefore, you cannot recover the database by using journal files created before you used either the Initialize_Tsns qualifier or the Set_Tsn qualifier in a restore-only-root operation. You should set the TSN to a value equal to or greater than the value that was in the lost .rdb file. If the TSN is set to a lower value than the value stored in the lost database root file, the database is corrupted, and it might return incorrect data or result in application failures. If the number you have selected is less than the Next CSN and Next TSN values, you will receive a fatal error message as follows: %RMU-F-VALLSSMIN, value (0:40) is less than minimum allowed value (0:74) for Set_Tsn=tsn After the set TSN and reinitialize TSN operations complete, and after you have verified the .rdb file, enabled after-image journaling, and the new .aij file is created, all .aij records are based on the new starting TSN and CSN numbers in the .rdb file. Although Oracle Corporation recommends that your backup strategy ensures that you maintain a current full and complete database backup file, it is possible to restore the database from current full by-area backup files only. This is accomplished by restoring the root and specifying the Noupdate_Files and Noset_ Tsn qualifiers. When you specify the Noset_Tsn qualifier, the TSN and CSN values on the restored database will be the same as those recorded in the backup file. When you specify the Noupdate_ Files qualifier, the database root is restored but RMU Restore Only_Root will not link that restored root to any of the area files, nor will it create or update the snapshot (.snp) files. By specifying the Noupdate_Files and Noset_Tsn qualifiers with the RMU Restore Only_Root command, you can use the following strategy to restore your database: 1. Restore the root from the most recent full by-area backup file. 2. Restore the storage areas by applying the by-area backup files in reverse order to their creation date. Apply the most recent by-area backup file first and the oldest by-area backup file last. (Be sure you do not restore any area more than once.) 3. Recover the database by applying the after-image journal (.aij) files. You can recover the .aij files manually by using the RMU Recover command. Or, if the state of your .aij files permits it, you can allow RMU Restore Only_Root to automatically recover the .aij files by not specifying the Norecovery qualifier with the last RMU Restore command you issue. For details on the automatic recovery feature of the RMU Restore command, see the help entry for the RMU Restore command. (The automatic recovery feature is not available for the RMU Restore Only_Root command.) When you use this strategy, be sure that the first RMU Restore command after the RMU Restore Only_Root command includes the most recent RDB$SYSTEM storage area. The RDB$SYSTEM storage area contains the structures needed to restore the other database storage areas. For this reason, Oracle Corporation suggests that you back up the RDB$SYSTEM storage area in every by-area backup operation you perform. See Example 6 in the Examples help entry under this command for a demonstration of this method. Note that the database backup file must be recent-differences between the database and backup file must be known, and the number of storage areas must be unchanged since the backup file was created. If you have moved a storage area, use the File qualifier to show its new location and the Snapshot qualifier to indicate the current version of the area's .snp file. NOTE You must perform a full and complete backup operation on your database when the RMU Restore Only_Root command completes. Oracle Corporation recommends that you define a new after-image journal configuration with the RMU Restore Only_Root command by using either the After_Journal or the Aij_Options qualifier. This action ensures that the new .aij file can be rolled forward in the event that another database restore operation becomes necessary. 4 Format (B)0RMU/Restore/Only_Root backup-file-spec [storage-area-list] Command Qualifiers x Defaults x /Active_IO=max-reads x /Active IO=3 /[No]After_Journal=file-spec x See description /[No]Aij_Options=journal-opts x See description /Directory=directory-spec x See description /[No]Initialize_Tsns x /Noinitialize_Tsns /Label=(label-name-list) x See description /Librarian[=options] x None /[No]Log x Current DCL verify value /[No]Media_Loader x See description /[No]New_Snapshots x /Nonew_Snapshots /Nodes_Max=number-cluster-nodes x Existing value /Options=file-spec x None /[No]Rewind x /Norewind /Root=root-file-spec x Existing value /[No]Set_Tsn=Tsn=n,Csn=m) x See description /Transaction_Mode=(modes-list) x /Transaction_Mode=Current /[No]Update_Files x /Update_Files /Users_Max=number-users x Existing value (B)0 File or Area Qualifiers x Defaults x /[No]Blocks_Per_Page=integer x /Noblocks_Per_Page /File=file-spec x See description /Read_Only x Current value /Read_Write x Current value /Snapshot=(Allocation=n,File=file-spec)x See description /[No]Spams x Current value /Thresholds=(val1[,val2[,val3]]) x Existing area file value 4 Parameters 5 backup-file-spec A file specification for the backup file produced by a previous RMU Backup command. The default file extension is .rbf. Note that you cannot perform a remote restore operation on an .rbf file that has been backed up to tape and then copied to disk. When copying .rbf files to disk from tape, be sure to copy them onto the system on which you will be restoring them. Depending on whether you are performing a restore operation from magnetic tape, disk, or multiple disks, the backup file specification should be specified as follows: o Restoring from magnetic tape If you used multiple tape drives to create the backup file, the backup-file-spec parameter must be provided with (and only with) the first tape drive name. Additional tape drive names must be separated from the first and subsequent tape drive names with commas, as shown in the following example: $ RMU/RESTORE /REWIND $111$MUA0:PERS_FULL_NOV30.RBF,$112$MUA1: o Restoring from multiple or single disk files If you used multiple disk files to create the backup file, the backup-file-spec parameter must be provided with (and only with) the first disk device name. Additional disk device names must be separated from the first and subsequent disk device names with commas. You must include the Disk_file qualifier. For example: RMU/RESTORE/ONLY_ROOT/DISK_FILE DISK1:[DIR1]MFP.RBF,DISK2:[DIR2], DISK3:[DIR3] As an alternative to listing the disk device names on the command line (which can exceed the line-limit lenght for a command line if you use several devices), you can specify an options file in place of the backup-file-spec. For example: $ RMU/RESTORE/ONLY-ROOT/DISK_FILE" @DEVICES.OPT" The contents of devices.opt might appear as follows: DISK1:[DIR1]MFP.RBF DISK2:[DIR2] DIS3:[DIR3] The backup files referenced from sjuch an options file are: DISK1:[DIR1]MFP.RBF DISK2:[DIR2]MFP01.RBF DISK3:[DIR3]MFP02.RBF 5 storage-area-list This option is a list of storage area names from the database. Use it in the following situations: o When you need to change the values for thresholds with the Thresholds qualifier or blocks per page with the Blocks_Per_ Page qualifier o When you need to change the names or version numbers specified with the Snapshot or the File qualifier for the restored database To use the storage-area-list option, specify the storage area name, not the system file name for the storage area. By restoring the database root only, you save the additional time normally needed to restore all the storage areas. Place commas between each storage area name in the list. If the storage area parameters have changed since the file was last backed up, the storage-area-list option updates the .rdb file parameters so they agree with the current storage area parameters in terms of location and file version. 4 Command_Qualifiers 5 Active_IO Active_IO=max-reads Specifies the maximum number of read operations to the backup file that the RMU Restore Only_Root command will attempt simultaneously. The value of the Active_IO qualifier can range from 1 to 5. The default value is 3. 5 After_Journal After_Journal=file-spec Noafter_Journal NOTE This qualifier is maintained for compatibility with versions of Oracle Rdb prior to Version 6.0. You might find it more useful to specify the Aij_Options qualifier, unless you are only interested in creating extensible .aij files. Specifies how RMU Restore Only_Root is to handle after-image journaling and .aij file creation, using the following rules: o If you specify the After_Journal qualifier and provide a file specification, RMU Restore Only_Root creates a new extensible .aij file and enables journaling. o If you specify the After_Journal qualifier but you do not provide a file specification, RMU Restore Only_Root creates a new extensible .aij file with the same name as the journal that was active at the time of the backup operation. o If you specify the Noafter_Journal qualifier, RMU Restore Only_Root disables after-image journaling and does not create a new .aij file. Note that if you specify the Noafter_Journal qualifier, there will be a gap in the sequence of .aij files. For example, suppose your database has .aij file sequence number 1 when you back it up. If you issue an RMU Restore Only_Root command with the Noafter qualifier, the .aij file sequence number will be changed to 2. This means that you cannot (and do not want to) apply the original .aij file to the restored database (doing so would result in a sequence mismatch). o If you do not specify an After_Journal, Noafter_Journal, Aij_ Options, or Noaij_Options qualifier, RMU Restore Only_Root recovers the journal state (enabled or disabled) and tries to reuse the .aij file or files. If you choose this option, take great care to either set the database root TSN and CSN correctly, or create a full and complete backup file of the database. Failure to do so might make it impossible for you to recover your database from the .aij file should it become necessary. However, if the .aij file or files are not available (for example, they have been backed up), after-image journaling is disabled. You cannot use the After_Journal qualifier to create fixed-size .aij files; use the Aij_Options qualifier. 5 Aij_Options Aij_Options=journal-opts Noaij_Options Specifies how RMU Restore Only_Root is to handle after-image journaling and .aij file creation, using the following rules: o If you specify the Aij_Options qualifier and provide a journal-opts file, RMU Restore Only_Root enables journaling and creates the .aij file or files you specify for the restored database. If only one .aij file is created for the restored database, it will be an extensible .aij file. If two or more .aij files are created for the database copy, they will be fixed-size .aij files (as long as at least two .aij files are always available). o If you specify the Aij_Options qualifier, but do not provide a journal-opts file, RMU Restore Only_Root disables journaling and does not create any new .aij files. o If you specify the Noaij_Options qualifier, RMU Restore Only_ Root disables journaling and does not create any new .aij files. o If you do not specify an After_Journal, Noafter_Journal, Aij_ Options, or Noaij_Options qualifier, RMU Restore Only_Root recovers the journaling state (enabled or disabled) and tries to reuse the .aij file or files. If you choose this option, take great care to either set the database root TSN and CSN correctly, or create a full and complete backup file of the database. Failure to do so might make it impossible for you to recover your database from the .aij file should it become necessary. However, if the .aij file or files are not available (for example, they have been backed up), after-image journaling is disabled. See Show After_Journal for information on the format of a journal-opts-file. 5 Directory Directory=directory-spec Specifies the default directory for the database root and the default directory for where the root can expect to find the database storage areas and snapshot files. See the Usage Notes for information on how this qualifier interacts with the Root, File, and Snapshot qualifiers and for warnings regarding restoring database files into a directory owned by a resource identifier. 5 Initialize_Tsns Initialize_Tsns Noinitialize_Tsns Initializes all transaction sequence number (TSN) values for the entire database by setting the values to zero. Each time a transaction is initiated against a database, a TSN is issued. The numbers are incremented sequentially over the life of the database. TSN and CSN values are each contained in a quadword with the following decimal format: high longword : low longword The high longword can hold a maximum user value of 32768 (215) and the low longword can hold a maximum user value of 4,294,967,295 (232). A portion of the high longword is used by Oracle Rdb for overhead. When you specify a TSN or CSN, you can omit the high longword and the colon if the TSN or CSN fits in the low longword. For example 0:444 and 444 are both valid input values. As your next TSN value approaches the maximum value allowed, you should initialize the TSNs. You can determine the next TSN and next commit sequence number (CSN) values by dumping the database root file, using the RMU Dump command with the Header and Option=Debug qualifiers. The Initialize_Tsns qualifier takes much more time to execute because all TSN values in the database are set to zero, which requires writing to every page in the database. When the database TSNs are reset, using the Initialize_Tsns qualifier, you should use the After_Journal qualifier or the Aij_Options qualifier and immediately perform a full database backup operation and create a new .aij file. This ensures continuity of journaling and the ability to recover the database. The default Noinitialize_Tsns qualifier does not initialize the database TSNs. Note that you cannot use the Initialize_Tsns with the Set_Tsn or Noset_Tsn qualifier in the same command. This restriction is required because Initialize_Tsns directs RMU Restore Only_Root to reset the TSN value to zero, while Set_Tsn directs RMU Restore Only_Root to reset the TSN to the value you have indicated, and Noset_Tsn leaves the TSN value unchanged. Never use the Initialize_Tsns qualifier if Replication Option for Rdb transfers have been defined for the database. The Initialize_Tsns qualifier does not reset the Replication Option for Rdb transfers. 5 Label Label=(label-name-list) Specifies the 1- to 6-character string with which the volumes of the backup file have been labeled. The Label qualifier is applicable only to tape volumes. You must specify one or more label names when you use the Label qualifier. You can specify a list of tape labels for multiple tapes. If you list multiple tape label names, separate the names with commas, and enclose the list of names within parentheses. In a normal restore operation, the Label qualifier you specify with the RMU Restore Only_Root command should be the same Label qualifier you specified with the RMU Backup command you used to back up your database. The Label qualifier can be used with indirect file references. See the Indirect-Command-Files help entry for more information. 5 Librarian Librarian=options Use the Librarian qualifier to restore files from data archiving software applications that support the Oracle Media Management interface. The file name specified on the command line identifies the stream of data to be retrieved from the Librarian utility. If you supply a device specification or a version number it will be ignored. Oracle RMU supports retrieval using the Librarian qualifier only for data that has been previously stored by Oracle RMU using the Librarian qualifer. The Librarian qualifier accepts the following options: o Trace_file=file-specification The Librarian utility writes trace data to the specified file. o Level_Trace=n Use this option as a debugging tool to specify the level of trace data written by the Librarian utility. You can use a pre-determined value of 0, 1, or 2, or a higher value defined by the Librarian utility. The pre-determined values are : - Level 0 traces all error conditions. This is the default. - Level 1 traces the entry and exit from each Librarian function. - Level 2 traces the entry and exit from each Librarian function, the value of all function parameters, and the first 32 bytes of each read/write buffer, in hexadecimal. o Logical_Names=(logical_name=equivalence-value,...) You can use this option to specify a list of process logical names that the Librarian utility can use to specify catalogs or archives where Oracle Rdb backup files are stored, Librarian debug logical names, and so on. See the specific Librarian documentation for the definition of logical names. The list of process logical names is defined by Oracle RMU prior to the start of any Oracle RMU command that accesses the Librarian application. The following OpenVMS logical names must be defined for use with a Librarian utility before you execute an Oracle RMU backup or restore operation. Do not use the Logical_Names option provided with the Librarian qualifier to define these logical names. o RMU$LIBRARIAN_PATH This logical name must be defined so that the shareable Librarian image can be loaded and called by Oracle RMU backup and restore operations. The translation must include the file type (for example, .exe), and must not include a version number. The shareable Librarian image must be an installed (known) image. See the Librarian implementation documentation for the name and location of this image and how it should be installed. o RMU$DEBUG_SBT This logical name is not required. If it is defined, Oracle RMU will display debug tracing information messages from modules that make calls to the Librarian shareable image. You cannot use device specific qualifiers such as Rewind, Density, or Label with the Librarian qualifier because the Librarian utility handles the storage meda, not Oracle RMU. 5 Log Log Nolog Specifies whether the processing of the command is reported to SYS$OUTPUT. Specify the Log qualifier to request that the progress of the restore operation be written to SYS$OUTPUT and the Nolog qualifier to suppress this report. If you specify neither, the default is the current setting of the DCL verify switch. (The DCL SET VERIFY command controls the DCL verify switch.) 5 Media_Loader Media_Loader Nomedia_Loader Use the Media_Loader qualifier to specify that the tape device from which the backup file is being read has a loader or stacker. Use the Nomedia_Loader qualifier to specify that the tape device does not have a loader or stacker. By default, if a tape device has a loader or stacker, RMU Restore Only_Root should recognize this fact. However, occasionally RMU Restore Only_Root does not recognize that a tape device has a loader or stacker. Therefore, when the first tape has been read, RMU Restore Only_Root issues a request to the operator for the next tape, instead of requesting the next tape from the loader or stacker. Similarly, sometimes RMU Restore Only_Root behaves as though a tape device has a loader or stacker when actually it does not. If you find that RMU Restore Only_Root is not recognizing that your tape device has a loader or stacker, specify the Media_ Loader qualifier. If you find that RMU Restore Only_Root expects a loader or stacker when it should not, specify the Nomedia_ Loader qualifier. 5 New_Snapshots New_Snapshots Nonew_Snapshots Allows you to specify whether to create new snapshot (.snp) files as part of a Restore Only_Root operation. The default is the Nonew_Snapshots qualifier, which causes the command to initialize the existing .snp files. If you specify the New_Snapshots qualifier, the command creates and initializes new .snp files. When you specify the New_ Snapshots qualifier, you should either delete the existing .snp files before the restore operation or purge the .snp files afterwards. 5 Nodes_Max Nodes_Max=number-cluster-nodes Specifies a new upper limit on the number of VMScluster nodes from which users can access the restored database. The Nodes_Max qualifier will accept values between 1 and 96 VMScluster nodes. The actual maximum is the highest number of VMScluster nodes possible in the current version of OpenVMS. The default value is the limit defined for the database before it was backed up. 5 Options Options=file-spec Specifies the options file that contains storage area names, followed by the storage area qualifiers that you want applied to that storage area. You can direct RMU Restore Only_Root to create an options file for use with this qualifier by specifying the Restore_Options qualifier with the RMU Backup, RMU Dump, and RMU Dump Backup commands. See Backup Database, Dump Database, and Dump Backup_ File for details. If you create your own options file, do not separate the storage area names with commas. Instead, put each storage area name on a separate line in the file. The storage area qualifiers that you can include in the options file are: Blocks_Per_Page, File, Snapshot, and Thresholds. You can use the DCL line continuation character, a hyphen (-), or the comment character (!) in the options file. The default file extension is .opt. See Example 5 in the Examples help entry under this command. 5 Rewind Rewind Norewind Specifies whether the tape that contains the backup file will be rewound before processing begins. The Norewind qualifier, the default, causes the search for the backup file to begin at the current tape position. The Rewind and Norewind qualifiers are applicable only to tape devices. RMU Restore Only_Root returns an error message if you use these qualifiers and the device is not a tape device. 5 Root Root=root-file-spec Requests that the database root (.rdb) be restored to the specified location. See the Usage Notes for information on how this qualifier interacts with the Directory, File, and Snapshot qualifiers and for warnings regarding restoring database files into a directory owned by a resource identifier. The Root qualifier is only meaningful when used with a multifile database. 5 Set_Tsn Set_Tsn=(Tsn=n, Csn=m) Noset_Tsn The Set_Tsn qualifier sets the database transaction sequence number (TSN) and commit sequence number (CSN) to the specified values. The correct value can be extracted from the original .rdb file if it is still accessible, or from the last .aij file if one is available. If that fails, you can use a TSN value larger than the maximum number of transactions applied to the database since it was created, or since TSNs were last initialized. The TSN and CSN values do not have to be the same value. However, you need to choose new values that are greater than the last values assigned to a transaction. Set_Tsn values are expected to be multiples of eight. If you specify a value that is not a multiple of eight, RMU Restore Only_Root assigns the next highest value that is a multiple of eight. (For example, if you specify Set_Tsn=(Tsn=90, Csn=90), RMU Restore Only_Root assigns the Next TSN a value of 96.) The default value for the Set_Tsn qualifier is the TSN and CSN values stored in the backup file plus 1,000,000 when TSNs are not being initialized. The new TSN and CSN values for most database applications should be larger than the number of transactions committed since the database was last backed up. Set the TSN and CSN values higher than this default increment value plus the value in the backup file when needed. You can determine the next TSN and CSN values by dumping the .rdb file, using the Option=Debug qualifier. The TSN and CSN values are each contained in a quadword with the following decimal format: high longword : low longword The high longword can hold a maximum user value of 32768 (215) and the low longword can hold a maximum user value of 4,294,967,295 (232). A portion of the high longword is used by Oracle Rdb for overhead. When you specify a TSN or CSN, you can omit the high longword and the colon if the TSN fits in the low longword. For example 0:444 and 444 are both valid TSN input values. The Noset_Tsn qualifier specifies that the root will be restored with the same TSN state as was recorded in the backup file. When you use the Noset_Tsn qualifier in conjunction with the Noupdate_Files qualifier, you can use a backup strategy that uses recent by-area full backup files in place of a recent full and complete backup file of the entire database. See Example 6 in the Examples help entry under this command. Note that you cannot use the Initialize_Tsns with the Set_Tsn or Noset_Tsn qualifier in the same command. This restriction is required because Initialize_Tsns directs RMU Restore Only_Root to reset the TSN value to zero, while Set_Tsn directs RMU Restore Only_Root to reset the TSN to the value you have indicated, and Noset_Tsn leaves the TSN value unchanged. 5 Transaction_Mode=(mode-list) Transaction_Mode=(mode-list) Sets the allowable transaction modes for the database root file created by the restore operation. The mode-list can include one or more of the following transaction modes: o All - Enables all transaction modes o Current - Enables all transaction modes that are set for the source database. This is the default transaction mode. o None - Disables all transaction modes o [No]Batch_Update o [No]Read_Only o [No]Exclusive o [No]Exclusive_Read o [No]Exclusive_Write o [No]Protected o [No]Protected_Read o [No]Protected_Write o [No]Read_Write o [No]Shared o [No]Shared_Read o [No]Shared_Write If you specify more than one transaction mode in the mode-list, enclose the list in parenthesis and separate the transaction modes from one another with a comma. Note the following: o When you specify a negated transaction mode, for example Noexclusive_Write, it indicates that exclusive write is not an allowable access mode for the copied database. o If you specify the Shared, Exclusive, or Protected transaction mode, Oracle RMU assumes you are referring to both reading and writing in that transaction mode. o No mode is enabled unless you add that mode to the list, or you use the All option to enable all transaction modes. o You can list one transaction mode that enables or disables a particular mode followed by another that does the opposite. For example, Transaction_Mode=(Noshared_Write, Shared) is ambiguous because the first value disables Shared_Write access and the second value enables Shared_Write access. Oracle RMU resolves the ambiguity by first enabling the modes as specified in the modes-list and then disabling the modes as specified in the modes-list. The order of items in the list is irrelevant. In the example presented previously, Shared_Read is enabled and Shared_Write is disabled. 5 Update_Files Update_Files Noupdate_Files The Update_Files qualifier specifies that the root will be restored, and RMU Restore Only_Root will attempt to link that restored root to the area files. In addition, the snapshot (.snp) file will be updated or created. This is the default. The Noupdate_Files qualifier specifies that the restore operation will restore the root, but it will not link that restored root to any of the area files, nor will it create or update the .snp files. When you use the Noupdate_Files qualifier in conjunction with the Noset_Tsn qualifier, you can use a backup strategy that uses recent by-area full backup files in place of a recent full and complete backup file of the entire database. See Example 6 in the Examples help entry under this command 5 Users_Max Users_Max=number-users Specifies a new upper limit on the number of users that can simultaneously access the restored database. The valid range is between 1 and 2032 users. The default value is the value defined for the database before it was backed up. 4 File_or_Area_Qualifiers NOTE Use these qualifiers to reconcile the information in the database root file with the storage area files on disk. These values can get out of synchronization when changes have been made to storage areas or snapshot files after the backup from which you are restoring the database root file was created. Setting these parameters updates the data in the root file only; it does not change the attributes of the storage areas or snapshot files themselves. 5 Blocks_Per_Page Blocks_Per_Page=integer Noblocks_Per_Page Updates the database root file with the number of blocks per page for the storage area. Use this qualifier to update the root when the blocks per page for a storage area has changed since the backup file from which you are restoring was created. This qualifier does not change the page size of a storage area itself; its purpose is to update the database root file with corrected information. If you use the default, the Noblocks_Per_Page qualifier, RMU Restore Only_Root takes the page size for the storage area from the page size specified for the database you backed up. This is a positional qualifier. This qualifier conflicts with storage areas that have a uniform page format. 5 File File=file-spec Updates the database root file with the file specification for the storage-area-name parameter it qualifies. Use this qualifier to update the root when the file specification for a storage area has changed since the backup file from which you are restoring the root was created. (For example, if you have used the RMU Move_Area command since the backup file was created.) This qualifier does not change the file specification of the storage area it qualifies; its purpose is to update the database root file with corrected information. When you specify the File qualifier, you must supply a file name. See the Usage Notes for information on how this qualifier interacts with the Root, Snapshot, and Directory qualifiers. This qualifier is not valid for single-file databases. This is a positional qualifier. 5 Read_Only Updates the database root file to reflect the read-only attribute for the storage area it qualifies. Use this qualifier to update the root when the read/write or read-only attribute has changed since the backup file from which you are restoring has changed. This qualifier does not change the attribute of the storage area it qualifies; its purpose is to update the database root file with corrected information. If you do not specify the Read_Only or the Read_Write qualifier, the storage areas is restored with the read/write attributes that were in effect when the database was backed up. 5 Read_Write Updates the database root file to reflect the read/write attribute for the storage area it qualifies. Use this qualifier to update the root when the read/write or read-only attribute has changed since the backup file from which you are restoring has changed. This qualifier does not change the attribute of the storage area it qualifies; its purpose is to update the database root file with corrected information. If you do not specify the Read_Only or the Read_Write qualifier, the storage areas is restored with the read/write attributes that were in effect when the database was backed up. 5 Snapshot Snapshot=(Allocation=n,File=file-spec) Updates the database root file to reflect the snapshot allocation or snapshot file specification (or both) for the area it qualifies. Use this qualifier to update the root when the snapshot attributes have changed since the backup file from which you are restoring the database root has changed. This qualifier does not change the attributes of the snapshot file it qualifies; its purpose is to update the database root file with corrected information. See the Usage Notes for information on how this qualifier interacts with the Root, Snapshot, and Directory qualifiers. The Snapshot qualifier is a positional qualifier. When you do not specify the Snapshot qualifier, RMU Restore Only_ Root restores snapshot areas according to the information stored in the backup file. 5 Spams Spams Nospams Updates the database root file to reflect the space area management (SPAM) information for the storage areas in the storage-area-list. Use this qualifier when the setting of SPAM pages (enabled or disabled) has changed since the backup file from which you are restoring the root was created. This qualifier does not change the attributes of the storage area it qualifies; its purpose is to update the database root file with corrected information. Use the Spams qualifier to update the root file information to indicate that SPAM pages are enabled for the storage areas qualified; use the Nospams qualifier to update the root file information to indicate that SPAM pages are disabled for the storage areas qualified. The default is to leave the attribute unchanged from the setting recorded in the backup file. This is a positional qualifier. 5 Thresholds Thresholds=(val1[,val2[,val3]]) Updates the database root file to reflect the threshold information for the storage areas in the storage-area-list. Use this qualifier when the threshold values have changed since the backup file from which you are restoring the root was created. This qualifier does not change the attributes of the storage area it qualifies; its purpose is to update the database root file with corrected information. This is a positional qualifier. The Thresholds qualifier applies only to storage areas with a mixed page format. If you do not use the Thresholds qualifier with the RMU Restore Only_Root command, Oracle Rdb uses the storage area's thresholds as recorded in the backup file. See the Oracle Rdb7 Guide to Database Performance and Tuning for more information on SPAM thresholds. 4 Usage_Notes o To use the RMU Restore Only_Root command for a database, you must have the RMU$RESTORE privilege in the root file access control list (ACL) for the database or the OpenVMS SYSPRV or BYPASS privilege. o The RMU Restore Only_Root command provides two qualifiers, Directory, and Root, that allow you to specify the target for the restored database root file. In addition, the Directory, File, and Snapshot file qualifiers allow you to specify a target for updates to the database root for the storage area and snapshot file locations. The target can be just a directory, just a file name, or a directory and file name. If you use all or some of these qualifiers, apply them as follows: - Use the Root qualifier to indicate the target for the restored database root file. - Use local application of the File qualifier to specify the current location of a storage area file if its location has changed since the database was backed up. The storage area is not affected by this qualifier. This qualifier updates the location of the storage area as recorded in the database root file. - Use local application of the Snapshots qualifier to specify the current location of a snapshot file if its location has changed since the database was backed up. The snapshot file is not affected by this qualifier. This qualifier updates the location of the snapshot file as recorded in the database root file. - Use the Directory qualifier to specify a default target directory for the root file and as a default directory for where the storage areas and snapshot files currently reside. The default target directory is where the database root file is restored if a directory specification is not specified with the Root qualifier. The default directory for the storage area and snapshot files is the directory specification with which the root file is updated if these files are not qualified with the Root, File, or Snapshot qualifier. It is also the default directory with which the Root file is updated for files qualified with the Root, File, or Snapshot qualifier if these qualifiers do not include a directory specification. Note the following when using these qualifiers: - Global application of the File qualifier when the target specification includes a file name causes RMU Restore Only_ Root to update the file name recorded in the database root file for all storage areas to be the same file name. - Global application of the Snapshot qualifier when the target specification includes a file name causes RMU Restore Only_Root to update the file name recorded in the database root file for all snapshot files to be the same file name. - Specifying a file name or extension with the Directory qualifier is permitted, but causes RMU Restore Only_Root to restore the database root file to the named directory and file and update the file name recorded in the database root file for all the storage areas and snapshot files to be the same directory and file specification. o When you restore a database root into a directory owned by a resource identifier, the ACE for the directory is applied to the database root file ACL first, and then the Oracle RMU ACE is added. This method is employed to prevent database users from overriding OpenVMS file security. However, this can result in a database which you consider yours, but to which you have no Oracle RMU privileges to access. See the Oracle Rdb Guide to Database Maintenance for details. o Only the database parameter values and the storage area parameter values for which there are qualifiers can be updated in the database root (.rdb) file using the restore-only-root operation. All other database and storage area parameter values that have changed since the database was last backed up must be reapplied to the .rdb file using the SQL ALTER DATABASE statement. o There are no restrictions on the use of the Nospams qualifier option with storage areas that have a mixed page format, but the use of the Nospams qualifier typically causes severe performance degradation. The Nospams qualifier is useful only where updates are rare and batched, and access is primarily by database key (dbkey). o You must set both TSN and CSN values at the same time. You cannot set the TSN value lower than the CSN value; however, you can set a CSN value higher than the TSN value. o The RMU Restore Only_Root command cannot be used if any storage area has been extended since the backup operation was done. You can use the RMU Dump Backup command with the Option=Root qualifier to determine if this is the case. 4 Examples Example 1 To prevent corruption of your databases, check your CSN and TSN values and set them to zero based on when they approach the maximum. First, enter an RMU Dump command to display the next CSN and next TSN values: $ RMU/DUMP/HEADER=(SEQUENCE_NUMBERS) MF_PERSONNEL . . . Sequence Numbers... - Transaction sequence number Next number is 0:256 Group size is 0:32 - Commit sequence number Next number is 0:256 Group size is 0:32 If the next CSN and the next TSN values are approaching the maximum number allowed, you must perform the following operations to initialize all TSN and CSN values to the value zero in your database. The operation might take some time to execute as it writes to every page in the database. First, create a backup file for the database. Then restore the database and initialize the CSN and TSN values with the Initialize_Tsns qualifier. Then, enter an RMU Dump command again to examine the next CSN and next TSN values. This example shows that both values have been set to zero. If you displayed the database pages, you would also notice that all TSN and CSN values are set to zero. $ RMU/BACKUP MF_PERSONNEL MF_PER_124.RBF $ RMU/RESTORE/ONLY_ROOT /INITIALIZE_TSNS MF_PER_124.RBF $ RMU/DUMP/HEADER=(SEQUENCE_NUMBERS) MF_PERSONNEL . . . Sequence Numbers... - Transaction sequence number Next number is 0:0 Group size is 0:32 - Commit sequence number Next number is 0:0 Group size is 0:32 Example 2 Perform the following to set the TSN and CSN values to a number that you select; a number that is greater than or equal to the next CSN and next TSN values. If the number you have selected is less than the next CSN and next TSN values recorded in the database header, you receive an error as follows: $ RMU/RESTORE/ONLY_ROOT/SET_TSN=(TSN=40,CSN=40) _$ MF_PERSONNEL.RBF %RMU-F-TSNLSSMIN, value (0:40) is less than minimum allowed value (0:224) for /SET_TSN=TSN %RMU-F-FTL_RSTR, Fatal error for RESTORE operation at 18-JUN-1997 16:59:19.32 Enter a number equal to or greater than the next CSN and next TSN values recorded in the database header: $ RMU/RESTORE/ONLY_ROOT/SET_TSN=(TSN=274,CSN=274) - _$ MF_PERSONNEL.RBF Enter an RMU Dump command to see the next CSN and next TSN values: $ RMU/DUMP/HEADER=(SEQUENCE_NUMBERS) MF_PERSONNEL . . . Sequence Numbers... - Transaction sequence number Next number is 0:288 Group size is 0:32 - Commit sequence number Next number is 0:288 Group size is 0:32 - Database bind sequence number Next number is 0:288 Example 3 The following RMU Restore Only_Root command restores the database root file from the database backup file (.rbf) to another device: $ RMU/RESTORE/ONLY_ROOT/ROOT=DXXV9:[BIGLER.TESTING]MF_PERSONNEL - _$ MF_PERSONNEL_BACKUP.RBF The following DIRECTORY command confirms that the MF_ PERSONNEL.RDB file was restored in the specified directory: $ DIRECTORY DXXV9:[BIGLER.TESTING]MF_PERSONNEL.RDB Directory DXXV9:[BIGLER.TESTING] MF_PERSONNEL.RDB;1 21-JAN-1991 14:37:36.87 Total of 1 file. Example 4 Use the File=file-spec qualifier to update the .rdb file with a storage area's new location. If you have moved a storage area to a new location, use the File qualifier to show its new location and the Snapshot qualifier to indicate the current version of the area's snapshot (.snp) file. Enter the following RMU commands to execute a series of operations that use the File and Snapshot qualifiers in a restore-only-root operation to update the .rdb file with new information since the database was last backed up. Back up the database file: $ RMU/BACKUP MF_PERSONNEL MFPERS_122.RBF. Move the area to another directory: $ RMU/MOVE_AREA MF_PERSONNEL JOBS - _$ /FILE=[BIGLER.MFTEST.TEST1]JOBS.RDA With the RMU Restore Only_Root command, give the area name, and specify both the storage area file specification and its new location. Also specify the snapshot (.snp) file with its correct version. Note that .snp file version numbers increment with the RMU Move_Area command. $ RMU/RESTORE/ONLY_ROOT MFPERS_122.RBF JOBS - _$ /FILE=[BIGLER.MFTEST.TEST1]JOBS.RDA - _$ /SNAPSHOT=(FILE=[BIGLER.V41MFTEST]JOBS.SNP;2) Display the .rdb file header and note that the file is correctly updated. The dump of the database root file lists these file specifications: $ RMU/DUMP/HEADER MF_PERSONNEL DXXV9:[BIGLER.MFTEST.TEST1]JOBS.RDA;1 DXXV9:[BIGLER.MFTEST]JOBS.SNP;2 Verify the .rdb file to be certain that it has been properly and completely updated relative to the files and their version numbers that comprise the database. $ RMU/VERIFY/ROOT MF_PERSONNEL Example 5 The following command achieves the same results as the RMU Restore Only_Root command in Example 4, but uses an options file to specify the current location of the JOBS storage area and the associated .snp file. $ RMU/RESTORE/ONLY_ROOT MFPERS_122.RBF - _$ JOBS/OPTIONS=OPTIONS_FILE.OPT $ ! $ TYPE OPTIONS_FILE.OPT JOBS /FILE=[BIGLER.V41MFTEST.TEST1]JOBS.RDA - /SNAPSHOT=(FILE=BIGLER.V41MFTEST]JOBS.SNP) Example 6 The following example demonstrates the use of the Noset_Tsn qualifier and the Noupdate_Files qualifier to restore a database using by-area backup files. In addition, it demonstrates the automatic recovery feature of the RMU Restore command. $ ! $ SET DEFAULT DISK1:[USER] $ ! $ ! Create .aij files for the database. Because three .aij files are $ ! created, fixed-size after-image journaling will be used. $ ! $ RMU/SET AFTER_JOURNAL/ENABLE/RESERVE=4 - _$ /ADD=(name=AIJ1, FILE=DISK2:[CORP]AIJ_ONE) - _$ /ADD=(name=AIJ2, FILE=DISK2:[CORP]AIJ_TWO) - _$ /ADD=(NAME=AIJ3, FILE=DISK2:[CORP]AIJ_THREE) - _$ MF_PERSONNEL %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery $ ! $ ! $ ! For the purposes of this example, assume the backup operation $ ! recommended in the preceding warning message is done, but $ ! that the time between this backup operation and the following $ ! operations is several months so that this backup file is too $ ! old to use in an efficient restore operation. $ ! $ ! Update the DEPARTMENTS table. $ ! $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL'; SQL> -- SQL> -- On Monday, insert a new row in the DEPARTMENTS table. The SQL> -- new row is stored in the DEPARTMENTS storage area. SQL> -- SQL> INSERT INTO DEPARTMENTS cont> (DEPARTMENT_CODE, DEPARTMENT_NAME, MANAGER_ID, cont> BUDGET_PROJECTED, BUDGET_ACTUAL) cont> VALUES ('WLNS', 'Wellness Center', '00188', 0, 0); 1 row inserted SQL> SQL> COMMIT; SQL> DISCONNECT DEFAULT; SQL> EXIT $ ! $ ! Perform a by-area backup operation, including half of the $ ! storage areas from the mf_personnel database. $ ! $ RMU/BACKUP/INCLUDE=(RDB$SYSTEM, EMPIDS_LOW, EMPIDS_MID, - _$ EMPIDS_OVER, DEPARTMENTS) MF_PERSONNEL - _$ DISK3:[BACKUP]MONDAY_FULL.RBF %RMU-I-NOTALLARE, Not all areas will be included in this backup file $ ! $ ! Update the SALARY_HISTORY table. $ ! $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL'; SQL> -- SQL> -- On Tuesday, one row is updated in the SQL> -- SALARY_HISTORY storage area. SQL> -- SQL> UPDATE SALARY_HISTORY cont> SET SALARY_END ='20-JUL-1993 00:00:00.00' cont> WHERE SALARY_START='14-JAN-1983 00:00:00.00' cont> AND EMPLOYEE_ID = '00164'; 1 row updated SQL> COMMIT; SQL> DISCONNECT DEFAULT; SQL> EXIT $ ! $ ! On Tuesday, back up the other half of the storage areas. $ ! $ RMU/BACKUP/INCLUDE=(SALARY_HISTORY, JOBS, EMP_INFO, - _$ MF_PERS_SEGSTR, RDB$SYSTEM) MF_PERSONNEL - _$ DISK3:[BACKUP]TUESDAY_FULL.RBF %RMU-I-NOTALLARE, Not all areas will be included in this backup file $ ! $ ! On Wednesday, perform additional updates. $ ! $ SQL SQL> ATTACH 'FILENAME MF_PERSONNEL'; SQL> -- SQL> -- Update another row in the SALARY_HISTORY table: SQL> UPDATE SALARY_HISTORY cont> SET SALARY_START ='23-SEP-1991 00:00:00.00' cont> WHERE SALARY_START='21-SEP-1981 00:00:00.00' cont> AND EMPLOYEE_ID = '00164'; 1 row updated SQL> COMMIT; SQL> DISCONNECT DEFAULT; SQL> EXIT $ ! $ ! Assume the database is lost on Wednesday. $ ! $ ! Restore the database root from the latest full-area backup file. $ ! $ RMU/RESTORE/ONLY_ROOT/NOUPDATE_FILES/NOSET_TSN - _$ DISK3:[BACKUP]TUESDAY_FULL.RBF/LOG %RMU-I-AIJRSTBEG, restoring after-image journal "state" information %RMU-I-AIJRSTJRN, restoring journal "AIJ1" information %RMU-I-AIJRSTSEQ, journal sequence number is "0" %RMU-I-AIJRSTSUC, journal "AIJ1" successfully restored from file "DISK2:[CORP]AIJ_ONE.AIJ;1" %RMU-I-AIJRSTJRN, restoring journal "AIJ2" information %RMU-I-AIJRSTNMD, journal has not yet been modified %RMU-I-AIJRSTSUC, journal "AIJ2" successfully restored from file "DISK2:[CORP]AIJ_TWO.AIJ;1" %RMU-I-AIJRSTJRN, restoring journal "AIJ3" information %RMU-I-AIJRSTNMD, journal has not yet been modified %RMU-I-AIJRSTSUC, journal "AIJ3" successfully restored from file "DISK2:[CORP]AIJ_THREE.AIJ;1" %RMU-I-AIJRSTEND, after-image journal "state" restoration complete %RMU-I-RESTXT_00, Restored root file DISK1:[USER]MF_PERSONNEL.RDB;1 %RMU-I-AIJRECBEG, recovering after-image journal "state" information %RMU-I-AIJRSTAVL, 3 after-image journals available for use %RMU-I-AIJRSTMOD, 1 after-image journal marked as "modified" %RMU-I-LOGMODSTR, activated after-image journal "AIJ2" %RMU-I-AIJISON, after-image journaling has been enabled %RMU-W-DOFULLBCK, full database backup should be done to ensure future recovery %RMU-I-AIJRECEND, after-image journal "state" recovery complete $ ! $ ! Restore the database areas, starting with the most recent $ ! full-area backup file. (If the RDB$SYSTEM area is not in the $ ! most recent full-area backup file, however, it must be restored $ ! first.) Do not restore any area more than once. $ ! $ ! Specify the Norecovery qualifier since there are additional $ ! backup files to apply. $ ! $ RMU/RESTORE/AREA/NOCDD/NORECOVER - _$ DISK3:[BACKUP]TUESDAY_FULL.RBF - _$ RDB$SYSTEM, SALARY_HISTORY, JOBS, - _$ EMP_INFO, MF_PERS_SEGSTR/LOG %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]MF_PERS_DEFAULT.RDA;1 at 18-JUN-1997 16:14:40.88 %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]SALARY_HISTORY.RDA;1 at 18-JUN-1997 16:14:41.28 %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]JOBS.RDA;1 at 18-JUN-1997 16:14:41.83 %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]EMP_INFO.RDA;1 at 18-JUN-1997 16:14:42.06 %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]MF_PERS_SEGSTR.RDA;1 at 18-JUN-1997 16:14:42.27 %RMU-I-RESTXT_24, Completed full restore of storage area DISK1:[USER]JOBS.RDA;1 at 18-JUN-1997 16:14:42.49 %RMU-I-RESTXT_24, Completed full restore of storage area DISK1:[USER]EMP_INFO.RDA;1 at 18-JUN-1997 16:14:42.74 . . . %RMU-I-RESTXT_01, Initialized snapshot file DISK1:[USER]MF_PERS_DEFAULT.SNP;1 %RMU-I-LOGINIFIL, contains 100 pages, each page is 2 blocks long %RMU-I-RESTXT_01, Initialized snapshot file DISK1:[USER]EMP_INFO.SNP;1 %RMU-I-LOGINIFIL, contains 100 pages, each page is 2 blocks long . . . %RMU-I-AIJWASON, AIJ journaling was active when the database was backed up %RMU-I-AIJRECFUL, Recovery of the entire database starts with AIJ file sequence 0 %RMU-I-COMPLETED, RESTORE operation completed at 18-JUN-1997 16:14:46.82 $ ! $ ! Complete restoring database areas by applying the most $ ! recent full-area backup file. However, do not include $ ! the RDB$SYSTEM table because that was already restored $ ! in the previous restore operation. This restore $ ! operation will attempt an automatic recovery of the .aij files. $ ! $ RMU/RESTORE/AREA/NOCDD DISK3:[BACKUP]MONDAY_FULL.RBF - _$ EMPIDS_LOW, EMPIDS_MID, EMPIDS_OVER, DEPARTMENTS/LOG %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]EMPIDS_OVER.RDA;1 at 18-JUN-1997 16:20:05.08 %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]EMPIDS_MID.RDA;1 at 18-JUN-1997 16:20:05.40 %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]EMPIDS_LOW.RDA;1 at 18-JUN-1997 16:20:05.91 %RMU-I-RESTXT_21, Starting full restore of storage area DISK1:[USER]DEPARTMENTS.RDA;1 at 18-JUN-1997 16:20:06.01 %RMU-I-RESTXT_24, Completed full restore of storage area DISK1:[USER]EMPIDS_OVER.RDA;1 at 18-JUN-1997 16:20:06.24 . . . %RMU-I-RESTXT_01, Initialized snapshot file DISK1:[USER]DEPARTMENTS.SNP;1 %RMU-I-LOGINIFIL, contains 100 pages, each page is 2 blocks long %RMU-I-RESTXT_01, Initialized snapshot file DISK1:[USER]EMPIDS_LOW.SNP;1 %RMU-I-LOGINIFIL, contains 100 pages, each page is 2 blocks long . . . %RMU-I-AIJWASON, AIJ journaling was active when the database was backed up %RMU-I-AIJRECFUL, Recovery of the entire database starts with AIJ file sequence 0 %RMU-I-AIJRECARE, Recovery of area DEPARTMENTS starts with AIJ file sequence 0 %RMU-I-AIJRECARE, Recovery of area EMPIDS_LOW starts with AIJ file sequence 0 %RMU-I-AIJRECARE, Recovery of area EMPIDS_MID starts with AIJ file sequence 0 %RMU-I-AIJRECARE, Recovery of area EMPIDS_OVER starts with AIJ file sequence 0 %RMU-I-AIJBADAREA, inconsistent storage area DISK1:[USER]DEPARTMENTS.RDA;1 needs AIJ sequence number 0 %RMU-I-AIJBADAREA, inconsistent storage area DISK1:[USER]EMPIDS_LOW.RDA;1 needs AIJ sequence number 0 . . . %RMU-I-LOGRECDB, recovering database file DISK1:[USER]MF_PERSONNEL.RDB;1 %RMU-I-AIJAUTOREC, starting automatic after-image journal recovery %RMU-I-LOGOPNAIJ, opened journal file DISK2:[CORP]AIJ_ONE.AIJ;1 %RMU-I-AIJONEDONE, AIJ file sequence 0 roll-forward operations completed %RMU-I-LOGRECOVR, 1 transaction committed %RMU-I-LOGRECOVR, 0 transactions rolled back %RMU-I-LOGRECOVR, 2 transactions ignored %RMU-I-AIJNOACTIVE, there are no active transactions %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJALLDONE, after-image journal roll-forward operations completed %RMU-I-LOGSUMMARY, total 1 transaction committed %RMU-I-LOGSUMMARY, total 0 transactions rolled back %RMU-I-LOGSUMMARY, total 2 transactions ignored %RMU-I-AIJSUCCES, database recovery completed successfully %RMU-I-AIJGOODAREA, storage area DISK1:[USER]DEPARTMENTS.RDA;1 is now consistent %RMU-I-AIJGOODAREA, storage area DISK1:[USER]EMPIDS_LOW.RDA;1 is now consistent %RMU-I-AIJGOODAREA, storage area DISK1:[USER]EMPIDS_MID.RDA;1 is now consistent . . . %RMU-I-AIJFNLSEQ, to start another AIJ file recovery, the sequence number needed will be 0 %RMU-I-COMPLETED, RESTORE operation completed at 18-JUN-1997 16:20:11.45 $ ! $ ! The database is now restored and recovered. However, if $ ! for some reason the automatic .aij file recovery was not $ ! possible (for example, if you had backed up the .aij files), $ ! apply the .aij files in the same order in $ ! which they were created. That is, if .aij files were backed $ ! up each night, apply aij_mon.aij first and aij_tues.aij second. Example 7 The following example demonstrates the use of the Directory, File, and Root qualifiers. First, the database is backed up, then a couple storage area files and a snapshot file are moved. The restore-only-root operation does the following: o The default directory is specified as DISK2:[DIR]. o The target directory and file name for the database root file is specified with the Root qualifier. The target directory specified with the Root qualifier overrides the default directory specified with the Directory qualifier. Thus, the RMU Restore Only_Root process restores the database root in DISK3:[ROOT] and names it COPYRDB.RDB. o The target directory for the EMPIDS_MID storage area is DISK4:[FILE]. The RMU Restore Only_Root process updates the database root file to indicate that EMPIDS_MID currently resides in DISK4:[FILE]. o The target for the EMPIDS_MID snapshot file is DISK5:[SNAP]EMPIDS_MID.SNP Thus, the RMU Restore Only_ Root process updates the database root file to indicate that the EMPIDS_MID snapshot file currently resides in DISK5:[SNAP]EMPIDS_MID.SNP. o The target file name for the EMPIDS_LOW storage area is EMPIDS. Thus, the RMU Restore Only_Root process updates the database root file to indicate that the EMPIDS_LOW storage area currently resides in the DISK2 default directory (specified with the Directory qualifier), and the file is currently named EMPIDS.RDA. o The target for the EMPIDS_LOW snapshot file is DISK5:[SNAP]EMPIDS.SNP. Thus, the RMU Restore Only_ Root process updates the database root file to indicate that the EMPIDS_LOW snapshot file currently resides in DISK5:[SNAP]EMPIDS.SNP. o Data for all the other storage area files and snapshot files remain unchanged in the database root file. $ ! Back up the database: $ ! $ RMU/BACKUP MF_PERSONNEL.RDB MF_PERSONNEL.RBF $ ! $ ! Move a couple of storage areas and a snapshot file: $ ! $ RMU/MOVE_AREA MF_PERSONNEL.RDB - _$ /DIRECTORY=DISK2:[DIR] - _$ EMPIDS_MID/FILE=DISK4:[FILE] - _$ /SNAPSHOT=(FILE=DISK3:[SNAP]EMPIDS_MID.SNP), - _$ EMPIDS_LOW/FILE=EMPIDS - _$ /SNAPSHOT=(FILE=DISK5:[SNAP]EMPIDS.SNP) $ ! $ ! Database root is lost. Restore the root and update the $ ! locations of the moved storage areas and snapshot file as $ ! recorded in the database root file because the locations $ ! recorded in the backup file from which the root is restored $ ! are not up-to-date: $ ! $ RMU/RESTORE/ONLY_ROOT MF_PERSONNEL.RBF - _$ /ROOT=DISK3:[ROOT]MF_PERSONNEL.RDB - _$ EMPIDS_MID/FILE=DISK4:[FILE] - _$ /SNAPSHOT=(FILE=DISK2:[DIR]EMPIDS_MID.SNP), - _$ EMPIDS_LOW/FILE=DISK2:[DIR]EMPIDS - _$ /SNAPSHOT=(FILE=DISK5:[SNAP]EMPIDS.SNP) 2 Server_After_Journal There are three RMU Server After_Journal commands, as follows: o The RMU Server After_Journal Start command starts the AIJ log server (ALS). o The RMU Server After_Journal Stop command stops the ALS. o The RMU Server After_Journal Reopen_Output command allows you to close and reopen the output file specified with the RMU Server After_Journal Start command. 3 Reopen_Output Allows you to close the current AIJ log server (ALS) output file for the specified database and open a new one. This allows you to see the current contents of the original ALS output file. 4 Description The RMU Server After_Journal Reopen_Output command allows you to reopen an ALS output file that was previously created with an RMU Server After_Journal Start command with the Output qualifier. (The ALS output file is opened for exclusive access by the ALS process.) Reopening the output file results in the current output file being closed and a new output file being created. The new output file has the same file name as the original output file, but its version number is incremented by one. The ALS is an optional process that flushes log data to the after-image journal (.aij) file. All database servers deposit transaction log data in a cache located in the database global section. If the ALS is active, it continuously flushes the log data to disk. Otherwise, server processes might block temporarily if the cache in the global section is full. 4 Format (B)0RMU/Server After_Journal Reopen_Output root-file-spec 4 Parameters 5 root-file-spec Specifies the database root file for which you want to reopen the ALS output file. 4 Usage_Notes o To use the RMU Server After_Journal Reopen_Output command for a database, you must have RMU$OPEN privilege in the root file access control list (ACL) for the database or the OpenVMS WORLD privilege. o To issue the RMU Server After_Journal Reopen_Output command successfully, the database must be opened. Other users can be attached to the database when this command is issued. o To determine whether the ALS is running, use the RMU Show Users command. 4 Examples Example 1 In the following example the first Oracle RMU command starts the log server and specifies an output file. The second Oracle RMU command reopens the ALS output file, so you can view the data that is contained in the ALS output file so far. $ RMU/SERVER AFTER_JOURNAL START MF_PERSONNEL/OUT=ALS $ ! Database updates occur $ RMU/SERVER AFTER_JOURNAL REOPEN_OUTPUT MF_PERSONNEL $ ! View the ALS.OUT;-1 file: $ TYPE ALS.OUT;-1 -------------------------------------------------------------------- 16-OCT-1995 13:02:05.21 - Oracle Rdb V7.0-00 database utility started --------------------------------------------------------------------- . . . 3 Start Allows you to manually start the AIJ log server (ALS) for the specified database and specify a file for the AIJ log server output. 4 Description The ALS is an optional process that flushes log data to the after-image journal (.aij) file. All database servers deposit transaction log data in a cache located in the database global section. If the ALS is active, it continuously flushes the log data to disk. Otherwise, server processes might block temporarily if the cache in the global section is full. The ALS should be started only when AIJ processing is a bottleneck. Typically, multiuser databases with medium to high update activity can benefit from using the ALS. You can start the ALS either manually, using the RMU Server After_Journal Start command, or automatically when the database is opened (by specifying LOG SERVER IS AUTOMATIC in the SQL ALTER DATABASE command). By default, the ALS startup is set to manual. 4 Format (B)0RMU/Server After_Journal Start root-file-spec Command Qualifier x Default x /Output=file-spec x See description 4 Parameters 5 root-file-spec Specifies the database root file for which you want to start the ALS. 4 Command_Qualifiers 5 Output Output=file-spec Specifies the file for the ALS output file. Use this qualifier in anticipation of issuing an RMU Server After_Journal Reopen_ Output command. By specifying the output file, you will know the location of, and therefore can view, the ALS output file. By default, the ALS output file is not available to the user. 4 Usage_Notes o To use the RMU Server After_Journal Start command for a database, you must have RMU$OPEN privilege in the root file access control list (ACL) for the database or the OpenVMS WORLD privilege. o The ALS can be started only if the database is open and if after-image journaling is enabled. o The RMU Server After_Journal Start command can be issued while users are attached to the database. o If the ALS process stops abnormally, regardless of whether the current setting of the ALS is automatic or manual, the only way to restart it is to use the RMU Server After_Journal Start command. o To determine whether the ALS is running, use the RMU Show Users command. o Any errors encountered when you