PI08783: WMQ Z/OS: A UOW FOR A SHARED QUEUE IS NOT BACKED OUT CORRECTLY WHEN IT IS SHUNTED AND THE QUEUE MANAGER IS RESTARTED

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• When an indoubt unit of work ( UOW ) containing only shared
  queue work is shunted and then the queue manager is restarted,
  the UOW is not backed out correctly. This leaves uncommitted
  MQPUTs and MQGETs for the shared queue.
  
  Symptoms include:
  - The CURDEPTH for the shared queue is non-zero, but a browse
    of the queue results in RC 2033 MQRC_NO_MSG_AVAILABLE.
    DISPLAY QSTATUS has UNCOM(YES).
  
  - RESOLVE CHANNEL with COMMIT or BACKOUT does not clear the
    messages from the queue.
  
  - DELETE QLOCAL for the queue results in:
    CSQM110I CSQMUQLC QLOCAL(<queue>) QSGDISP(SHARED)
     HAS INCOMPLETE UNITS OF RECOVERY
    CSQM090E CSQMUQLC FAILURE REASON CODE X'00D44005'
    CSQ9023E CSQMUQLC ' DELETE QLOCAL' ABNORMAL COMPLETION
  
  - The MVS command
      SETXCF FORCE,CONNECTION,STRNAME=nnnnCSQ_ADMIN,
      CONNAME=<connection>
    where "nnnn" is the Queue Sharing Group ( QSG ) name, results
    in the message:
  
    IXC354I THE SETXCF FORCE REQUEST FOR CONNECTION <connection>
    IN STRUCTURE nnnnCSQ_ADMIN WAS REJECTED:
    FORCE CONNECTION NOT ALLOWED FOR PERSISTENT LOCK OR
    SERIALIZED  LIST
  

Local fix

• If a Unit of Work is in this condition, contact the Support
  Center for a job to clear the uncommitted messages.
  

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: Messages in an indoubt unit of work     *
  *                      containing only activity against shared *
  *                      queues are left uncommitted when the    *
  *                      unit of work is resolved, if the unit   *
  *                      of work was shunted after the last      *
  *                      checkpoint occurred, and the queue      *
  *                      manager was restarted.                  *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  During queue manager restart the state of indoubt units of work
  is rebuilt, and a record is kept of any resource managers with
  an interest. During this process shunted INSERT UOWD records are
  ignored, however if the unit of work contained only shared queue
  operations this can lead to the RM interest not being recorded
  correctly. When the unit of work is subsequently resolved (e.g.
  automatically during RRS resync, or manually using RESOLVE
  INDOUBT), the lack of interest means the correct RM is not
  invoked for commit/backout, leading to the shared queue messages
  remaining uncommitted.
  

Problem conclusion

• CSQRSHUN is changed to no longer flag shunted INSERT UOWD
  records to be ignored, so that the correct RM interest can be
  determined when the indoubt state is rebuilt during queue
  manager restart.
  010Y
  100Y
  CSQJPPLG
  CSQRRCSL
  CSQRSHUN
  

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:

  UI14280 UI14281

Modules/Macros

• CSQJPPLG CSQRRCSL CSQRSHUN
  

Fix information

• Fixed component name

  WMQ Z/OS V7

• Fixed component ID

  5655R3600

Applicable component levels

• R010 PSY UI14280

     UP14/02/05 P F402 Ž

• R100 PSY UI14281

     UP14/02/05 P F402 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.