PI27399: OUT OF SYNCPOINT PUBLISHES DO NOT TERMINATE NEWLY CREATED NESTED UNITS OF WORK WHEN BACKING OUT A FAILED PUBLISH

Obtain the fix for this APAR.

APAR status

• Closed as unreproducible in next release.

Error description

• When an out of syncpoint publish operation starts a new UOW,
  but the publish fails, the messages put inside the publish are
  backed out; however the (now empty) UOW is not ended. On
  subsequent publish operations this UOW is reused and will not
  be ended until the task ends (for long running tasks such as
  channels this leads to an increase in the number of log records
  in the unit of work that must be shunted and a consequent
  increase in the amount of data logged).
  .
  Addtional symptoms:
  CSQJ160I Long-running UOW found
  CSQR026I Long-running UOW shunted
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of WebSphere MQ for z/OS Version 7 *
  *                 Release 1 Modification 0.                    *
  ****************************************************************
  * PROBLEM DESCRIPTION: Out of syncpoint publishes, that create *
  *                      a nested unit of work, such as a        *
  *                      message that meets the topic delivery   *
  *                      criteria (NPMSGDLV/PMSGDLV) for all     *
  *                      subscribers, that fail to successfully  *
  *                      complete do not terminate the           *
  *                      associated UR.                          *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  Publish/Subscribe based work may result in a nested unit of work
  being created. These are created when a message meets the
  criteria for the topic attributes NPMSGDLV and PMSGDLV, which
  may result in the put to the topic failing if the message fails
  to be delivered to subscriptions specified in the attribute. If
  the work fails to complete successfully, such as the case of a
  failed put, the nested UOW is backed out, but in the case where
  a new UR was created to handle the nested UOW, this is not
  terminated. This results in the UR remaining present, resulting
  in additional persistent work reusing this UR.
  
  This can result in the UR being considered long running, and
  being shunted. If a large number of messages reuse this UR, the
  UR may result in the logs being filled at a higher rate, due to
  more space being required to shunt the log records.
  

Problem conclusion

Temporary fix

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

Comments

• Nested unit of work processing has been updated to detect if a
  nested UOW is required, in the case where another UR is
  currently active. A nested UOW will only be created when
  correctly required.
  

APAR Information

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

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

  PI27881 UI23432

Modules/Macros

• CSQERSAV CSQIDEL3 CSQIRSAV CSQLRSAV CSQMRSAV
  CSQMSUSB CSQMTPUT
  

Fix information

• Fixed component name

  WMQ Z/OS V7

• Fixed component ID

  5655R3600

Applicable component levels

• R100 PSY UI23432

     UP15/01/16 P F501 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.