PM91419: WMQ Z/OS: UNCOMMITTED MESSAGES REMAIN ON A SHARED QUEUE AFTER A FAILURE DURING PUBLISH PROCESSING.

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Messages published to shared queues have a QSTATUS with
  UNCOM(YES) but no handle details.
  .
  There was a failure during publish processing (in this case, it
  was the abend in PM79003).   If any abend occurs during publish
  processing, published messages may not be subsequently
  committed or backed out as requested by the application
  program.
  
  Additional Symptom(s) Search Keyword(s):
  backout back out commit pubsub pub/sub
  

Local fix

• None besides resolving the prior abend.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of WebSphere MQ for z/OS Version 7 *
  *                 Release 0 Modification 1 and Release 1       *
  *                 Modification 0.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: Following an abend during publish       *
  *                      processing, various errors can occur    *
  *                      including:                              *
  *                      - Messages only delivered to some       *
  *                        subscribers despite NPMSGDLV/PMSGDLV  *
  *                        being set to ALL or ALLDUR.           *
  *                      - Messages on shared queues left        *
  *                        uncommitted after MQCMIT/MQBACK       *
  *                      - MQRC_NO_MSG_AVAILABLE when getting    *
  *                        messages from indexed queues          *
  *                      - Abend 5C6-00C90600 when getting       *
  *                        messages from indexed queues          *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  When publishing to a topic with NPMSGDLV/PMSGDLV set to ALL or
  ALLAVAIL, a nested unit of work is created to allow the publish
  operation to be rolled back independently of the current unit of
  work.
  When the nested unit of work is created, state information for
  the current unit of work is stored, ready to be restored when
  the nested unit of work completes.
  If an abend occurs while the nested unit of work is active,
  CSQMTPUT does not correctly restore the state of the original
  unit of work, leading to the reported problems occurring for
  messages put or got in the original unit of work prior to the
  abending publish operation.
  

Problem conclusion

• Recovery processing in CSQMTPUT is changed to roll back the
  nested unit of work and restore the state of the original unit
  of work when handling an abend during publish processing.
  010Y
  100Y
  CSQMTPUT
  

Temporary fix

• *********
  * HIPER *
  *********
  

Comments

APAR Information

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

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

  UK97570 UK97571

Modules/Macros

• CSQMTPUT
  

Fix information

• Fixed component name

  WMQ Z/OS V7

• Fixed component ID

  5655R3600

Applicable component levels

• R010 PSY UK97570

     UP13/10/16 P F310 Ž

• R100 PSY UK97571

     UP13/10/16 P F310 Ž

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.