IBM Support

IV31569: "RE-REGISTRATION OF THE MESSAGE RECOVERY THREAD ONLY IF SETTINGS HAVE CHANGED"

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • "re-registration of the message recovery thread only if settings
    have changed"
    .
    "Currently, every time a Process Server application target JVM
    is re-)started, a re-registration of the Work Manager message
    recovery thread with the BPEScheduler happens (CScopeProvidedBea
    n.reRegisterCleanup()). This requires the cancellation and re-
    creation of a scheduler entry through scheduler APIs. In a dis-
    tributed platform high availability environment or on a multi-
    servant setup on zOS re-registrations can happen multiple
    times (redundantly), due to the number of JVM starts.
    However, generally the re-registration is only required, if the
    configuration settings for th recovery thread have changed.
    have changed. The fix will introduce this conditional re-
    registration, thus greatly reducing the number of re-
    registrations."
    

Local fix

Problem summary

  • Problem Description:
    When using WorkManager based navigation, on server restart in a
    clustered environment, the following assertion
    may be observed in SystemOut.log:
    
     Assertion violation !(Cannot schedule Message Cleanup) in
    method >>
    com.ibm.bpe.framework.navigation.CScopeProviderBean.reRegisterCl
    eanup(CScopeProviderBean.java:198)
    
    Problem Summary:
    Currently, every time a Process Server application target
    JVM is (re-)started, a re-registration of the Work Manager
    message
    recovery thread with the BPEScheduler happens
    (CScopeProviderBean.reRegisterCleanup()). This requires the
    cancellation
    and re-creation of a scheduler entry through scheduler APIs. In
    a
    distributed platform high availability environment or on a
    multi-servant
    setup on Z/Os re-registrations can happen multiple times
    (redundantly),
    due to the number of JVM starts. Under unfavorable
    circumstances,
    this can lead to deadlocks and unavailability of the Scheduler.
    
    Problem Conclusion:
    
    Re-registration is only required, if the configuration settings
    for the recovery thread
    have changed. The fix will introduce this conditional
    re-registration,
    thus greatly reducing the number of re-registrations."
    

Problem conclusion

  • Problem fixed
    iFix available on top V620 FP3
    

Temporary fix

Comments

APAR Information

  • APAR number

    IV31569

  • Reported component name

    BUS PRC CHOREOG

  • Reported component ID

    5655FLW11

  • Reported release

    620

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2012-11-07

  • Closed date

    2013-02-06

  • Last modified date

    2013-02-06

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

    PM76642

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

    JR45336 JR45346 JR45347 IV36142

Fix information

  • Fixed component name

    BUS PRC CHOREOG

  • Fixed component ID

    5655FLW11

Applicable component levels

[{"Business Unit":{"code":"BU048","label":"IBM Software"},"Product":{"code":"SS7J6S","label":"WebSphere Enterprise Service Bus"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"6.2","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
11 September 2024