JR30371: REPLICATION ENGINE FIXES FOR DB2 LUW V9.1 FP7

DB2 Version 9.1 Fix Pack 7  for Linux, UNIX and Windows
DB2 Version 9.1 Fix Pack 7a  for Linux, UNIX and Windows
DB2 Version 9.1 Fix Pack 8  for Linux, UNIX and Windows
DB2 Version 9.1 Fix Pack 9  for Linux, UNIX and Windows
DB2 Version 9.1 Fix Pack 10  for Linux, UNIX and Windows
DB2 Version 9.1 Fix Pack 11  for Linux, UNIX and Windows
DB2 Version 9.1 Fix Pack 12  for Linux, UNIX and Windows

APAR status

• Closed as new function.

Error description

Local fix

Problem summary

• The following Replication Server defects have been fixed
  in V91FP7:
  .
  363138: SQLAPPLY fails with ASN1016 when disable_refresh=1
    and loadx_t
  User Affected: SQL Replication
  Problem Description: SQLAPPLY fails with ASN1016 for
    disable_refresh=1, when the member has loadx_type=6 or
    MEMBER_STATE='D'
  Problem Summary: SQLAPPLY should not issue error message
    ASN1016 for members with loadx_type=6 or
    MEMBER_STATE='D'
  Problem Conclusion: SQLAPPLY will not detect setting of
    disable_refresh for members with loadx_type=6 or
    MEMBER_STATE='D'
  .
  379123: SQLAPPLY fails with SQL104N if target table is CCD
    and TARGET_STRUCTURE value is 9
  User Affected: SQL Replication users
  Problem Description: SQLAPPLY fails with SQL104N if target
    table is CCD and TARGET_STRUCTURE value is 9
  Problem Summary: Apply missed IBMSNAP_COLUMNS in insert
    statement for type 9 target
  Problem Conclusion: added IBMSNAP_COLUMNS for type 9
    target
  .
  365917: SQLCAPTURE (in DPF): Restart LSN for partitions
    not updated properly
  User Affected: SQLCAPTURE
  Problem Description: The restart LSN for each partition is
    updated: (1) asynchronously by the admin thread (which
    doesn't guarantee the update is done in the same commit
    interval).  (2) only one partition's restart LSN is
    updated per commit interval in a round-robin fashion
    (which leaves all other partitions very outdated and
    makes warmstart re-read hours of log data that has
    already been processed)
  Problem Conclusion: SQLCAPTURE will now update every row
    in IBMSNAP_PARTITIONINFO upon commit interval.
    Previously, it updated only one row each commit interval
  .
  373983: Log reader buffer size default of 200KB (on LUW)
    is too small for some rare DB2 log records
    (max log rec size 256KB)
  Problem Conclusion: Increased default LOGBUFSIZE from
    200KB to 256KB.
  .
  368060: CAPTURE ABENDS0C4 while REINIT/STOP is being
    processed.
  User Affected: QCAPTURE, SQLCAPTURE
  Problem Description: The worker thread terminates before
    the logreader terminates.
  Problem Summary: The worker/logreader communication is
    inadequate
  Problem Conclusion: The worker turns on the stop flag to
    signal the logreader to stop.
  .
  369811: Before values for LOB/XML should not be counted in
    max message
  User Affected: QCAPTURE
  Problem Description: QCAPTURE was counting the LOB and XML
    columns for before values in max message size estimation
  Problem Summary: QCAPTURE should set not count those
    columns because they are never sent
  Problem Conclusion: QCAPTURE is now excluding those
    columns in max message size estimate
  .
  367038: Check IMPLICITVALUE in SYSCAT.COLUMNS for original
    ADDCOL default value
  User Affected: QCAPTURE
  Problem Description: QCAPTURE was not using the
    IMPLICITVALUE column as the default value for the added
    column
  Problem Summary: QCAPTURE should query the SYSCAT.COLUMNS
    table for the IMPLICITVALUE of the added column
  Problem Conclusion: QCAPTURE is now sending the
    IMPLICITVALUE column as the default value for the added
    column
  .
  328559: SUPPRESS DELETES with search condition
  User Affected: QCAPTURE
  Problem Description: QCAPTURE did not suppress deletes
    after search condition evaluation
  Problem Summary: QCAPTURE should also suppress the deletes
    after search condition evaluation
  Problem Conclusion: QCAPTURE now suppresses the deletes
    after search condition evaluation
  .
  362819: Admin thread should always fetch the subname from
    the SUBS table
  User Affected: QCAPTURE
  Problem Description: The admin thread only checked the
    subscriptions in memory for subname to be used in
    inserting signals, sometimes the subscriptions are not
    in memory thus causing problems with the signals when
    processing admin messages
  Problem Summary: The admin thread should always query the
    IBMQREP_SUBS table for the subname to be used in
    inserting signals when processing admin messages
  Problem Conclusion: The admin thread now always queries
    the IBMQREP_SUBS table for the subname to be used in
    inserting signals when processing admin messages
  .
  371385: statement owner should be invalidated in
    getPrepredStmt() if handle is stolen.
  User Affected: QCAPTURE
  Problem Description: SQL statement handles were not
    invalidated when they were overtaken by new owners,
    causing erroneous SQL statements to be executed
  Problem Summary: Internal SQL statement cache manager
    should invalidate obsolete SQL statement handles
  Problem Conclusion: Obsolete SQL statement handles are
    now invalidated
  .
  373770: asnmon: bug fix for QCAPTURE/QAPPLY V8 monitor
    interval to be in seconds not in milliseconds
  User Affected: All users that use asnmon to monitor
    QReplication
  Problem Description: The monitor_interval for QCAPTURE and
    QAPPLY was changed in V9 to be in milliseconds.  Monitor
    code however processes the monitor_interval as if it was
    in seconds although the value is in millisecond.
  Problem Summary: Asnmon interprets the monitor_interval of
    V9 QCAPTURE/QAPPLY in seconds.
  Problem Conclusion: Asnmon had to be changed to interpret
    QCAPTURE/QAPPLY monitor_interval column in V8 in seconds
    and V9 in milliseconds.
  .
  373773: asntdiff: signed interger overflow when addressing
    storage in spill file > 2GB
  User Affected: Users that run asntdiff on a 64bit system,
    against tables with huge amount of data.
  Problem Description: Asntdiff runs into an integer
    overflow problem when it tries to access more than 2GB
    of data in the spill file.
  Problem Summary: Asntdiff uses a signed integer to address
    records in the spill file.
  Problem Conclusion: On 64bit systems, asntdiff is now
    capable to address spill file content beyond 2GB due to
    a 8Byte offsets.
  .
  366114: linuxamd64: ASN1980E "Asnpwd" : "" : "Initial".
    The program did not complete successfully because
    "Operation not permitted"
  User Affected: The problem was only observed during an
    asnpwd init invocation on a 64bit linux system.
  Problem Description: Asnpwd init invocation results in
    ASN1980E  "Asnpwd" : "" : "Initial". The program did not
    complete successfully because "Operation not permitted"
  Problem Summary: Asnpwd does not allow the user to create
    the initial password file, but returns an
    "Operation not permitted error".
  Problem Conclusion: The asnpwd file handling operations
    had to be changed.
  .
  

Problem conclusion

• The defects mentioned in this APAR have
  been fixed in DB2 LUW V91 FP7.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  REPLICATION SER

• Fixed component ID

  5724N9800

Applicable component levels

• R910 PSY

     UP