IBM Support

PI07866: THE ZOS MQ MASTER ADDRESS SPACE MIPS INCREASES BETWEEN QMGR RESTARTS

A fix is available

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • The customer is running MQ V7.0.1. They noticed the continual
    MIPs increase in the Queue Manager (QMGR) Master address space
    between QMGR restarts.
    
    They usually do not re-start z/OS queue managers unless there is
    an LPAR IPL or maintenance upgrade.  The dump of the QMGR showed
    a large amount of MTMEs and system trace showed many entries for
    CSQMNMTIN.
    
    The Change Team can now see the cause of the leak of
    MTME control blocks.  CSQMMTPR CSQM_UNMARK_MSG_TOKEN processing
    causes MTME_FINVALID to be set as a result of the destructive
    MQGET for the message (MTME_FINVALID being set means that the
    message can be deleted the next time the queue is processed by
    CSQMMTIN).
    
    However, the message is being browsed at the same
    time and CSQMMTPR is invoked for CSQM_MARK_MSG_TOKEN which
    results in MTME_FINVALID being reset.  This means that the MTME
    will never be released, causing the problem seen here.
    *
    Additional Keywords:
    CPU performance storage leak Subpool 229
    Key 7 SP229 KEY7
    ADDITIONAL SYMPTOM :
    ORPHANED MTME INDICATED BY ABEND5C6 00C90600 OR 00C90900 IN
    CSQIMGE3 WITHOUT INCREASE IN MIPS OR STORAGE USE
    

Local fix

Problem summary

  • ****************************************************************
    * USERS AFFECTED: All users of WebSphere MQ Version 7          *
    *                 Release 0 Modification 1 and                 *
    *                 Release 1 Modification 0.                    *
    ****************************************************************
    * PROBLEM DESCRIPTION: Increased CPU usage and creeping        *
    *                      storage usage detected in queue         *
    *                      managers with applications browsing and *
    *                      marking messages like WSAS JMS MDB      *
    *                      listeners, or applications using        *
    *                      message selection.                      *
    *                                                              *
    *                      The following abends 5C6-00C90600 in    *
    *                      CSQIMGE3 and 5C6-00C90B00 in CSQIMGE9   *
    *                      can be generated as result of this      *
    *                      problem.                                *
    ****************************************************************
    * RECOMMENDATION:                                              *
    ****************************************************************
    Applications browsing (MQGET with MQGMO_BROWSE_... options) and
    marking messages (MQGMO_MARK_BROWSE_... options) do not unmark
    messages selected as candidates for browsing when the browse
    process detects that the message has been deleted by a
    destructive MQGET (running in another thread) resulting in MTME
    (message token mark entry) records not freed causing the
    creeping storage usage.
    
    The abends in CSQIMGE3 and CSQIMGE9 are caused by CSQMMTIN
    updating cursors in the queue with the stale data found in
    the orphaned MTME records.
    

Problem conclusion

  • CSQIMGE3 and CSQIMGE1 have been changed to unmark message token
    entries (MTME) for messages deleted during the browse process
    before trying to get the next message in the queue.
    010Y
    100Y
    CSQEMTIN
    CSQIMGEG
    CSQIMGE1
    CSQIMGE3
    CSQIVALM
    CSQMGET
    CSQMMTIN
    CSQMMTIS
    CSQMMTPR
    

Temporary fix

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

Comments

APAR Information

  • APAR number

    PI07866

  • Reported component name

    WMQ Z/OS V7

  • Reported component ID

    5655R3600

  • Reported release

    010

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    YesHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2013-12-11

  • Closed date

    2014-03-31

  • Last modified date

    2015-10-16

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

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

    UI16580 UI16581

Modules/Macros

  • CSQEMTIN CSQIMGEG CSQIMGE1 CSQIMGE3 CSQIVALM CSQMGET  CSQMMTIN
    CSQMMTIS CSQMMTPR
    

Fix information

  • Fixed component name

    WMQ Z/OS V7

  • Fixed component ID

    5655R3600

Applicable component levels

  • R010 PSY UI16580

       UP14/04/23 P F404 ¢

  • R100 PSY UI16581

       UP14/04/23 P F404 ¢

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.

[{"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SG19M","label":"APARs - z\/OS environment"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"7.0.1","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
16 October 2015