PI17684: WMQ Z/OS: REDUCE IXLLSTC REQUEST(MONITOR_KEYRANGE) ACTION(STOP) CALLS WHEN AN MQCLOSE OCCURS FOR A SHARED TRIGGERED QUEUE

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• The customer had an issue with a build-up of DWXB control blocks
  representing kiss and tell QSG notifications.  Their
  configuration used shared application queues with trigger-every
  and a shared init queue.  The z/OS systems for the QSG were
  split over 2 sites, so some of the LPARs had a long distance to
  the CF hosting the application structure.  This resulted in CF
  requests having a significant response time.
  
  A large portion of the DWXB notifications were for EKTM_TAP_FLAG
  when an application queue was closed by the last connection with
  it open.  This notification resulted in a call to CSQESMO1 to
  stop keyrange monitoring for the application queue.
  
  It was noticed from the trace that we would call CSQESMO1, which
  resulted in an IXLLSTC REQUEST(MONITOR_KEYRANGE) ACTION(STOP)
  call, for every notification.  However, there was no evidence of
  any processing calling CSQEBMO1 to start this type of
  monitoring.  If we're not starting keyrange monitoring for a
  particular queue, then the stop monitoring call is redundant as
  it does not change the state of the monitoring.
  
  This APAR should improve the performance of this processing by
  removing unnecessary calls to IXLLSTC from CSQESMO1.
  
  Additional keywords:
  CSQEBMO1 CSQESMO1 IXLLSTC
  SHARED APPLICATION QUEUE
  TRIGTYPE TRIGGERTYPE EVERY
  DWXB
  EKTM_TAP_FLAG
  KAT Kiss and Tell
  

Local fix

• N/A
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of WebSphere MQ Version 7          *
  *                 Release 1 Modification 0.                    *
  ****************************************************************
  * PROBLEM DESCRIPTION: Users of triggered shared queues are    *
  *                      calling XES unnecessarily to stop list  *
  *                      transition monitoring when monitoring   *
  *                      has not been started.                   *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  When a triggered shared queue is closed and no other
  applications have the queue open for input in the QSG then all
  other active QSG members are notified of the queue close.
  
  When the queue uses trigger type EVERY, the receivers of this
  notification generate requests to stop monitoring list
  transitions even if this monitoring is not active.
  

Problem conclusion

• This is a performance enhancement APAR.
  
  CSQEBMON sets a flag when list transition monitoring is started.
  CSQESMON checks if list transition monitoring is active before
  calling XES to stop it.
  100Y
  CSQEBMON
  CSQESMON
  

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:

  UI18840

Modules/Macros

• CSQEBMON CSQESMON
  

Fix information

• Fixed component name

  WMQ Z/OS V7

• Fixed component ID

  5655R3600

Applicable component levels

• R100 PSY UI18840

     UP14/06/27 P F406 ¢

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.