IBM Support

PJ36094: PE PROCESSES CRASHING WITH EXCEPTION 0XC0000003 DURING STARTUP/S HUTDOWN

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • PMR 51397,SGC,724
    Description of Problem:
    During startup/shutdown of the Process Engine, some of the
    processes crash with the following exception:
    Application exception occurred:
     App: ../../bin\VWJs.exe (pid=5992)
     When: 8/9/2008 @ 23:11:18.866
     Exception number: 80000003 (hardcoded breakpoint)
    The process crashing is not always the same (there are
    occurences of killfnsw.exe, vwtime.exe, OCOR_Listen, ...).
    From the userdump provided, this seems to happen during
    process exit.
    Other exceptions seen in the Windows system logs are 0xC0000120
    and 0xC0000142 which happen during process initialization.
    Again, this can occur during startup or shutdown of the
    Process Engine, even after a clean reboot.  For example, in
    the logs attached, there is one occurence where MKF_shutdown
    crashes durning PE shutdown (see windows system logs at
    16:37:30 UTC+1).  CSS has not been able to establish a link
    with the 0xC0000003 exception and have requested the customer
    to capture a user dump using Microsoft UserDump when the
    problem reoccurs.
    Regarding exception 0xC0000142, the customer has already tuned
    the desktop heap size but this did not help.  It was initially
    set to 2MB, then reduced to 1MB and these exceptions are still
    occuring.
    The customer has however been able to gather DrWatson logs +
    user dump for the 0xC0000003 exception (attached to the
    Defect).
    How long has the problem been occurring (recent changes)? :
    This customer has been experiencing startup issues since at
    least P8PE-3.5.1.  The main issue was initially processes
    crashing during PE started with a 0xC0000005 rexception.  This
    has been fixed in Defect 751482.  Occurences of exceptions
    0xC0000142 have been seen intermittently but this has started
    to happen more often since the customer upgraded to 3.5.3.
    Additionally, when the problem starts to occur, it takes
    several attempts (including reboots) to get the PE up and
    running.
    The attached Excel file shows a time line of the events for
    8/9, 8/10.  In most occasions the PE is started by just
    starting the Process Services Manager, with PE and PPM set to
    start automatically.  At 1:15PM and 5:04PM, the PE is started
    manually, i.e. The IMS Control Service is started manually,
    then the isolated region is mounted with "vwtool -s" and the
    Process Manager Service is started last.  This was used as a
    workaround to defect 751482 and seems to help in this case too.
     This will however not help when the PE fails to stop.
    Does this occur on more than one station / server? :
    Only on production.
    Is there a workaround? :
    Partial workaround by mounting the isolated region manually,
    see above.
    What is the impact to the customer/system? :
    The whole server is recycled every night and given the issues,
    this cannot happen unattended and requires the intervention of
    a qualified P8 admin.  This doesn't help the customer's poor
    perception of the software reliability either.
    Can the problem be replicated on an internal system? NO
    NO
    Are there any similar or related Defects?
    751482, parent 384732
    What logs were collected and where are they located?
    elogs, Windows application and system logs, DrWatson logs +
    UserDump.
    Configuration/Environment :
    Server OS :  Windows 2000 SP4
    Database :   SQL Server 2000 SP4
    P8PE-3.5.3-001 + OneOff for 751482
    

Local fix

Problem summary

  • ****************************************************************
    * USERS AFFECTED:                                              *
    * Users using PE on Windows                                    *
    ****************************************************************
    * PROBLEM DESCRIPTION:                                         *
    * PE processes crashing intermittently with exception          *
    * 0xc0000003 during startup/shutdown.                          *
    * The process crashing is not always the same (there are       *
    * occurences of killfnsw.exe, vwtime.exe, OCOR_Listen, ...).   *
    * Other exceptions seen in the Windows system logs are         *
    * 0xC0000120 and 0xC0000142 which happen during process        *
    * initialization.                                              *
    * Again, this can occur during startup or shutdown of the      *
    * Process Engine, even after a clean reboot.                   *
    ****************************************************************
    * RECOMMENDATION:                                              *
    * Upgrade to PE4.5.1                                           *
    ****************************************************************
    

Problem conclusion

  • Fixed in PE4.5.1
    

Temporary fix

Comments

APAR Information

  • APAR number

    PJ36094

  • Reported component name

    PROCESS ENGINE

  • Reported component ID

    5724R7607

  • Reported release

    450

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2009-04-01

  • Closed date

    2010-11-08

  • Last modified date

    2010-11-08

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

    PJ35738

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

Fix information

  • Fixed component name

    PROCESS ENGINE

  • Fixed component ID

    5724R7607

Applicable component levels

  • R451 PSY

       UP

[{"Business Unit":{"code":"BU048","label":"IBM Software"},"Product":{"code":"SSCQTT4","label":"Process Engine"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"450","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
08 November 2010