JR38795: REPLICATION ENGINE FIXES FOR DB2 LUW V9.5 FP8

DB2 Version 9.5 Fix Pack 8 for Linux, UNIX, and Windows
DB2 Version 9.5 Fix Pack 9 for Linux, UNIX, and Windows
DB2 Version 9.5 Fix Pack 10 for Linux, UNIX, and Windows

APAR status

• Closed as program error.

Error description

Local fix

Problem summary

• .
  The following Replication Server engine fixes are included
  in DB2 LUW V9.5 FP8:
  .
  .
  825273: QAPPLY: SpillQ creation on Schema with load_phase
   I is not immediately committed on MQ causing Agents to
   fail with ASN0569E on MQOPEN
  .
  Problem Description: When starting a subscription with
   load phase, QAPPLY would create the SPILLQ but did not
   call MQCMIT immediately when committing subscription
   state in DB2.  This may race with QAPPLY agents
   attempting to spill row changes for the subscription
   getting failure in MQOPEN with ASN0569E reported.
  .
  Problem Conclusion: When starting a subscription with load
   phase, QAPPLY will immediately commit on both MQ and DB2
   after creating SPILLQ and updating subscription state so
   that agents attempting spilling will not fail in MQOPEN.
  .
  .
  821839: QAPPLY does not cleanup load msg, data and part
   files in DPF environment
  .
  Problem Description: In a DPF environment, QAPPLY does not
   cleanup all the temporary load data files that get
   created for each partition after doing initial load for a
   subscription.
  .
  Problem Conclusion: QAPPLY is now fixed to cleanup all the
   temporary load data files generated in a DPF environment
   after doing the initial load for a subscription.
  .
  .
  827892: DPF - no subtrans created for coordinator of
   global transaction with no prepare
  .
  Problem Description: In a DPF environment, the restart LSN
   for a partition might advance beyond the commit LSN of
   certain global transactions before the transaction is
   actually published to the queue, causing the transaction
   to be skipped on restart if QCAPTURE crashes at that
   point.  Also when the catalog partition is not the lowest
   partition id and it shares a commit timestamp with
   another partition of a lower partition id, the non
   catalog partition might be skipped when publishing.
  .
  Problem Conclusion: QCAPTURE will correctly maintain the
   restart LSNs and there is no risk of data loss on restart
   after a crash.  Also, QCAPTURE will take into
   consideration the node number of the catalog partition
   when evaluating commit records with duplicate timestamps,
   in order to ensure that each transaction on each
   partition that contains data to be replicate is always
   published.
  .
  .
  813925: ADDCOL signal handling problem in P2P
   configurations
  .
  Problem Description: In a 3-node P2P environment ADDCOL an
   signal inserted on a node may not be handled correctly
   leading to a rejection of the newly added column on one
   or more nodes when following the internal ADDCOL
   propagation protocol. This also causes the subscriptions
   to be in an inconsistent state across the P2P
   configuration regarding the number of subscribed columns.
  .
  Problem Conclusion: The problem has been fixed.  Once the
   ADDCOL signal for a subscription is inserted on one node
   it is propagated to the other participating nodes and
   processed correctly. Once the protocol has been completed
   the new column is added to all nodes.
  .
  .
  810021: SQLAPPLY - user requires "rollback" after SQL0919N
   occurs
  .
  Problem Description: SQLAPPLY hangs after it receives
   SQL0919N from a z/OS server.
  .
  Problem Conclusion: With this fix, SQLAPPLY issues a
   DB2 ROLLBACK to resume replication.
  .
  .
  813958: ASNMON: asnmail always return 0 no matter mail
   send is failed or not
  .
  Problem Description: ASNMON always shows the alert mail
   sent successfully even though mail failed to send
  .
  Problem Conclusion: ASNMON shows the correctly mail sent
   status.
  .
  .
  710509: ASNMON: sql for APPLY_SUBSINACT alert is not
   correct
  .
  Problem Description: The query used for APPLY_SUBSINACT
   alert condition is not correct.
  .
  Problem Conclusion: Correct the query used for
   APPLY_SUBSINACT.  ASNMON will now send alerts when any
   subscription set fails or becomes inactive.
  

Problem conclusion

•  Refer to the previous section.
  

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:

  IC75639

Fix information

• Fixed component name

  WS Q-REPLIC LUW

• Fixed component ID

  5724N9801

Applicable component levels

• R950 PSY

     UP