RDOHELP72.HLB  —  START_TRANSACTION, Format  distributed-transaction-flag
  (B)0distributed-transaction-flag =

  q> DISTRIBUTED_TRANSACTION qwqqqqqqqqqqqqqqqqqqq>qqqqqqqqqqqqqqqwq>
                              m> DISTRIBUTED_TID distributed-tid qj

1  –  DISTRIBUTED_TRANSACTION

    A distributed transaction is a transaction that uses more than
    one database handle. Examples of a distributed transaction
    include:

    o  a transaction that attaches more than once to a single Oracle Rdb
       database

    o  a transaction that attaches to two or more Oracle Rdb databases

    o  a transaction that attaches to more than one database
       management system (an Oracle Rdb database and a Oracle CODASYL
       DBMS database, for example)

    Distributed transactions are managed by the DECdtm services.
    DECdtm services assigns a unique distributed transaction
    identifier (distributed TID) to each distributed transaction
    when the distributed transaction is started.

    Use the DISTRIBUTED_TRANSACTION keyword by itself (without the
    DISTRIBUTED_TID distributed-tid clause) when you want to start a
    distributed transaction.

    For complete information on distributed transactions, see the
    Oracle Rdb Guide to Distributed Transactions.

2  –  DISTRIBUTED_TID

    A keyword followed by a host language variable (distributed-tid)
    in application programs. When you want the transaction you are
    starting to join a distributed transaction, use this clause to
    specify the distributed TID of the distributed transaction that
    you want your transaction to join.

    The distributed-tid is a host language variable. You use the
    distributed-tid variable to hold the value of the distributed TID
    that DECdtm services generates and returns to the application.
    DECdtm services uses the distributed TID to distinguish the
    databases involved in a particular distributed transaction.

    Note that if you want to start a distributed transaction, you
    should use only the DISTRIBUTED_TRANSACTION clause. However, if
    you want your application to start a transaction that will join a
    distributed transaction, then you must use both the DISTRIBUTED_
    TRANSACTION clause and the DISTRIBUTED_TID distributed-tid
    clause.

    Your application must explicitly call the SYS$START_TRANS system
    service and you must specify the DISTRIBUTED_TID distributed-
    tid clause if you want the transaction you are starting to join
    a distributed transaction. The distributed-tid variable is an
    octaword (16 bytes) that you should declare and initialize to
    zero at the beginning of your application.

    Do not use the DISTRIBUTED_TID clause in interactive RDO. It is
    valid only in RDBPRE and Callable RDO.

    For complete information on distributed transactions, see the
    Oracle Rdb Guide to Distributed Transactions.
Close Help