1  –  unmark_buffer_field

    This field is incremented by one each time a modified buffer is
    written back to disk. Its value should be equal to the sum of
    the 14 following fields. These fields as well as the SPAM page
    field provide further detail about buffer unmark activity. Write
    operations of buffers that contain SPAM pages are included in the
    total although they are also counted separately by the SPAM page
    field.

    Note that unmarking one buffer may entail more than one I/O.

2  –  __transaction_field

    This field is incremented by one for each modified buffer that
    is written back to disk as a result of a COMMIT or ROLLBACK
    statement.

3  –  __pool_overflow_field

    This field is incremented by one for each modified buffer that
    is written back to disk as a result of a request to read in a new
    page from disk.

4  –  blocking AST field

    This field is incremented by one for each modified buffer that
    is written back to disk in response to a blocking AST (that is, a
    page in the buffer was requested by another user).

5  –  __lock_quota_field

    This field is incremented by one for each modified buffer that is
    written back to disk due to a user reaching his lock quota.

6  –  __lock_conflict_field

    This field is incremented by one for each modified buffer that is
    written back to disk to reduce the possibility of a deadlock when
    Oracle Rdb discovers a lock conflict.

7  –  __user_unbind_field

    This field is incremented by one for each modified buffer that is
    written back to disk as a result of a user's detaching from the
    database. This includes database detaches that occur as part of
    an Oracle RMU operation.

8  –  __batch_rollback_field

    This field is incremented by one for each modified buffer that
    is written back to disk because of a failed or rolled back batch-
    update transaction.

9  –  __new_area_mode_field

    This field is incremented by one for each modified buffer that is
    written back to disk as a result of a physical area being readied
    in a new lock mode.

10  –  __larea_change_field

    This field is incremented by one for each modified buffer that
    is written back to disk as a part of creating, deleting, or
    extending a logical area.

11  –  __incr_backup_field

    This field is incremented by one for each modified buffer that
    is written back to disk to support the requirements of the
    incremental backup feature. These buffers are always SPAM page
    buffers.

12  –  no AIJ access field

    This field is incremented by one for each modified buffer that
    is written back to disk as a safety precaution after the failure
    of an I/O to the .aij file. The buffers are unmarked under such
    circumstances only when fast commit processing is enabled to
    ensure that all committed data is safely in the database.

13  –  __truncate_snaps_field

    This field is incremented by one for each modified buffer that
    is written back to disk as a part of an online snapshot file
    truncation.

14  –  __checkpoint_field

    This field is incremented by one for each modified buffer that is
    written back to disk as a result of a checkpoint.

15  –  AIJ backup field

    This field is incremented by one for each modified buffer that
    is written back to disk to optimize the performance of the quiet-
    point backup of the .aij file.