IBM Support

PI07566: WEBSPHERE MQ Z/OS 7.1.0:MQ BATCH DELETE PROCESSING IMPROVEMENT

A fix is available

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • Batch delete processing can either be triggered when  the batch
    delete buffer is full (341 messages), or a put to a shared
    queue fails due to a "structure full" event.
    However, the latter is only attempted on the queue manager
    issuing the MQPUT. Thus the call can  still fail if the CF is
    full due to entries on the uncommitted list headers for other
    queue managers in the QSG.
    Exposure to this issue is increased if the queue managers doing
    the  MQGET calls are not doing any MQPUT calls to that same
    structure, as entries will build up on the uncommitted list
    header, until the batch  delete buffer is full, without any
    MQPUT calls that could trigger early batch delete processing
    due to a "structure full" event.
    Having a system with an even distribution of MQPUT and MQGET
    calls to the CF structure will reduce, though not eliminate,
    exposure to the structure full issue. The value that triggers
    normal batch delete processing can be improved,  to avoid
    waiting for a full buffer in a scenario where the structure
    could run out of space earlier.
    Using the number of queue managers in the QSG, and the size of
    the CF structure, MQ should be able to adapt the limit to
    something more appropriate.
    

Local fix

  • N/A.
    

Problem summary

  • ****************************************************************
    * USERS AFFECTED: All users of WebSphere MQ for z/OS Version 7 *
    *                 Release 1 Modification 0.                    *
    ****************************************************************
    * PROBLEM DESCRIPTION: Unexpected conditions of CF Structure   *
    *                      full and MQPUT calls to shared queues   *
    *                      on that structure failing with MQRC     *
    *                      2192 (MQRC_STORAGE_MEDIUM_FULL) due to  *
    *                      used structure space that has not been  *
    *                      released by queue managers in the QSG.  *
    ****************************************************************
    * RECOMMENDATION:                                              *
    ****************************************************************
    When MQGET calls for a shared queue are issued with the
    MQGMO_NO_SYNCPOINT option, the message will be moved from the
    queue list header in the CF application structure to the
    uncommitted list header for the particular queue manager.
    The entry is not deleted immediately, but added to a list of
    entries to delete by batch at a later stage.
    Depending on the setup of the CF structure and the QSG, this can
    result in a situation where entries left on the uncommitted list
    headers fill up the structure before the batch delete is
    triggered. For instance a fairly small CF structure, with a
    larger number of queue managers in the QSG and larger messages
    on the shared queues can cause this to happen.
    

Problem conclusion

  • A service parm was added to enable a calculation of the batch
    delete limit for application structures when the queue managers
    first connects to the structure, taking into account the size of
    the CF structure, as well as the number of queue managers in the
    QSG.
    Please note that will reduce the amount of storage used in the
    CF, but increase the CF CPU usage, as calls to delete entries
    from the CF structure are issued more often.
    The service parm should only be used under guidance from IBM
    support.
    100Y
    CSQEBDEL
    

Temporary fix

Comments

APAR Information

  • APAR number

    PI07566

  • Reported component name

    WMQ Z/OS V7

  • Reported component ID

    5655R3600

  • Reported release

    100

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2013-12-06

  • Closed date

    2014-01-29

  • Last modified date

    2014-04-02

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

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

    UI14614

Modules/Macros

  • CSQEBDEL
    

Fix information

  • Fixed component name

    WMQ Z/OS V7

  • Fixed component ID

    5655R3600

Applicable component levels

  • R100 PSY UI14614

       UP14/03/04 P F403

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":"7.1","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
02 April 2014