IBM Support

PK25600: OUT OF MEMORY IN VERSION 6 DEPOYMENT MANAGER WITH VERSION 5 NODES

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • 6.0.2 and 6.1 Deployment Managers with 5.x nodes on them can
    experience out of memory errors in larger environments due to
    interop code which handles the translation of cluster data
    between 6.x versions and 5.x versions.  This interop code was
    not designed with large scale topologies in mind, so
    improvements were needed in handling only the needed data and
    notifications instead of all data and notifications.
    

Local fix

Problem summary

  • ****************************************************************
    * USERS AFFECTED: Webpshere Application Server version 6       *
    *                 users with large version 6 topologies        *
    *                 and/or users running version 5 and           *
    *                 version 6 mixed node clusters.               *
    ****************************************************************
    * PROBLEM DESCRIPTION: Out of Memory exceptions causing        *
    *                      crashes of Deployment Manager           *
    *                                                              *
    ****************************************************************
    * RECOMMENDATION:                                              *
    *                                                              *
    ****************************************************************
    In an mixed version 5 and version 6 node topology, the dmgr
    registers to the bulletin board on behalf of all version 5
    members, and there is code which translates all of the
    version 5 descriptions to version 6 and vice versa. In a
    reasonably sized cell this extra work can cause the deployment
    manager to crash due to out of memory errors because of the
    extra work processing the large numbers of notifications
    passed around because the cell is mixed.
    
    This particular fix will benefit all users, regardless of
    topology size, but for those customers running in the range
    of up to 50 Application Server processes, they are unlikely to
    hit Out of Memory problems, and the fix will be a CPU
    utilization and performance enhancement.  It is unknown where
    the line is on how large the topology must be before Out of
    Memory exceptions may start occurring, and is also depending
    on the number of version 6 nodes compared to version 5 nodes,
    but the effect will be greater the larger the topology is.
    The scenario in which the OutOfMemory was discovered had 350
    processes spread across 8-10 nodes, with one node at version 6
    and the rest at version 5.  The more nodes which are at
    version 6 (and the fewer at version 5), the lower the amount
    of work the DeploymentManager will have to do in order to
    bridge the data between the different versions, decreasing
    the chance there will be an out of memory error.
    

Problem conclusion

  • Optimization improvements were backported to version 6.0.2
    from version 6.1 for the WLM SleeperThreadPool for processing
    the notifications, and further improvements were made on both
    version 6.0.2 and version 6.1 to the interoperabiltiy code to
    limit the scope of the notifications it registers for, which
    eliminates a vast percentage of unneeded notifications from
    being sent around the cell.
    
    The fix for this APAR is currently targeted for inclusion
    in fixpack 6.0.2.13 and 6.1.0.1.
    Please refer to the recommended updates page for delivery
    information:
    http://www.ibm.com/support/docview.wss?rs=180&uid=swg27004980
    

Temporary fix

Comments

APAR Information

  • APAR number

    PK25600

  • Reported component name

    WEBS APP SERV N

  • Reported component ID

    5724H8800

  • Reported release

    60W

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2006-05-25

  • Closed date

    2006-06-19

  • Last modified date

    2006-08-16

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

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

Modules/Macros

  • WLM
    

Fix information

  • Fixed component name

    WEBS APP SERV N

  • Fixed component ID

    5724H8800

Applicable component levels

  • R60A PSY

       UP

  • R60H PSY

       UP

  • R60I PSY

       UP

  • R60P PSY

       UP

  • R60S PSY

       UP

  • R60W PSY

       UP

  • R60Z PSY

       UP

  • R61A PSY

       UP

  • R61H PSY

       UP

  • R61I PSY

       UP

  • R61P PSY

       UP

  • R61S PSY

       UP

  • R61W PSY

       UP

  • R61Z PSY

       UP

[{"Business Unit":{"code":"BU053","label":"Cloud & Data Platform"},"Product":{"code":"SSEQTP","label":"WebSphere Application Server"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"6.0","Line of Business":{"code":"LOB45","label":"Automation"}}]

Document Information

Modified date:
19 October 2021