IBM Support

IC65822: Messages are workload balanced incorrectly in a clustering setup using pipelined channels.

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • Messages are not spread evenly between two cluster destinations
    despite the fact that all workload balancing attributes have
    been set equally. Upon further investigation, it is found that
    some of the messages are being redistributed to a remote queue
    manager in the cluster, even once they had arrived across a
    cluster channel.
    

Local fix

Problem summary

  • ****************************************************************
    USERS AFFECTED:
    Any users of WebSphere MQ using pipelined channels in a
    cluster.
    
    Platforms affected:
    All Distributed (iSeries, all Unix and Windows)
    ****************************************************************
    PROBLEM SUMMARY:
    During channel initialisation, a flag is set on the thread
    servicing the channel to indicate that it is servicing a
    cluster receiver channel. This flag is accessible by the
    workload balancing algorithm and allows it to verify whether
    messages are incoming from a remote destination. This knowledge
    is used to prevent messages from being routed to remote
    destinations if it has already come across a cluster channel,
    thus preventing messages from hopping across multiple locations
    in the cluster. However, this flag was not being set in the
    secondary thread of a pipelined channel, and as a result, the
    workload balancing routine was unaware of the fact that
    messages being processed had already arrived from a remote
    destination. This was then leading to messages being routed
    repeatedly around the cluster, leading to unexpected
    distributions of messages on each individual destination.
    

Problem conclusion

  • WebSphere MQ was modified to ensure that the secondary thread
    responsible for communications across the dual pipelined
    channel was initialised correctly to reflect the fact that it
    was servicing messages on a cluster receiver channel. As such,
    the workload balancing algorithm running in the pipelined
    thread is now successfully able to tell that the message has
    already arrived from a remote queue manager, and consequently,
    removes all remote destinations from its list of eligible
    destinations. This prevents messages routed over cluster
    channels from hopping between multiple queue managers in the
    cluster.
    
    ---------------------------------------------------------------
    The fix is targeted for delivery in the following PTFs:
    
                       v7.0
    Platform           Fix Pack 7.0.1.3
    --------           --------------------
    Windows            tbc_p700_0_1_3
    AIX                tbc_p700_0_1_3
    HP-UX (PA-RISC)    tbc_p700_0_1_3
    HP-UX (Itanium)    tbc_p700_0_1_3
    Solaris (SPARC)    tbc_p700_0_1_3
    Solaris (x86-64)   tbc_p700_0_1_3
    iSeries            tbc_p700_0_1_3
    Linux (x86)        tbc_p700_0_1_3
    Linux (x86-64)     tbc_p700_0_1_3
    Linux (zSeries)    tbc_p700_0_1_3
    Linux (Power)      tbc_p700_0_1_3
    
                       v6.0
    Platform           Fix Pack 6.0.2.10
    --------           --------------------
    Windows            U200318
    AIX                U832558
    HP-UX (PA-RISC)    U833084
    HP-UX (Itanium)    U833088
    Solaris (SPARC)    U833085
    Solaris (x86-64)   U832557
    iSeries            tbc_p600_0_2_10
    Linux (x86)        U832545
    Linux (x86-64)     U833083
    Linux (zSeries)    U833089
    Linux (Power)      U833087
    Linux (s390x)      U833086
    
    The latest available maintenance can be obtained from
    'WebSphere MQ Recommended Fixes'
    http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg27006037
    
    If the maintenance level is not yet available, information on
    its planned availability can be found in 'WebSphere MQ
    Planned Maintenance Release Dates'
    http://www-1.ibm.com/support/docview.wss?rs=171&uid=swg27006309
    ---------------------------------------------------------------
    

Temporary fix

Comments

APAR Information

  • APAR number

    IC65822

  • Reported component name

    WMQ WINDOWS V7

  • Reported component ID

    5724H7220

  • Reported release

    701

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2010-01-26

  • Closed date

    2010-02-24

  • Last modified date

    2010-02-24

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

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

    PM08381

Fix information

  • Fixed component name

    WMQ WINDOWS V7

  • Fixed component ID

    5724H7220

Applicable component levels

  • R701 PSY

       UP

[{"Business Unit":{"code":"BU048","label":"IBM Software"},"Product":{"code":"SSCPQ63","label":"APAR \/ Maintenance"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"7.0.1","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
24 February 2010