IBM Support

IT22101: amqrmppa process crashes if SVRCONN channel is running with an exit channel statistics and async consume

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as unreproducible in next release.

Error description

  • With channel statistics enabled on a SVRCONN channel with a
    channel exit in place and using the
    async consume feature the amqrmppa process crashes with a memory
    exception.
    
    The issue does not occur if a channel exit is not in place and
    channel statistics are enabled.
    
    The issue also does not occur if a channel exit is in place and
    channel statistics are not enabled.
    
    Failure Data Capture (FDC) records will be generated and example
    FDCs on UNIX/LINUX platforms would present the following Probe
    ID, component, comment and call stack.
    
    Probe Id: XC130003
    Component: xehExceptionHandler
    Comment1: SIGSEGV: address not mapped(0x64)
    
    O/S Call Stack for current thread:
    
    /opt/mqm/lib64/libmqmcs_r.so(xcsPrintStackForCurrentThread+0x3e)
    71b8bce]
    /opt/mqm/lib64/libmqmcs_r.so(+0x11537a)
    /opt/mqm/lib64/libmqmcs_r.so(PrepareDumpAreas+0xc07)
    /opt/mqm/lib64/libmqmcs_r.so(xcsFFSTFn+0xb75)
    /opt/mqm/lib64/libmqmcs_r.so(xehExceptionHandler+0x705)
    /usr/lib64/libpthread.so.0(+0xf5e0)
    /opt/mqm/lib64/libmqmr_r.so(rriCallSndExit+0xb9c
    /opt/mqm/lib64/libmqmr_r.so(rstSendAsyncMessage+0x411)
    /opt/mqm/lib64/libmqmr_r.so(rstConsumer+0xbe4)
    /opt/mqm/lib64/libmqzi_r.so(zstUserCallback+0x23a)
    /opt/mqm/lib64/libmqzi_r.so(zstAsyncConsume+0x1786)
    /opt/mqm/lib64/libmqzi_r.so(zstAsyncConsumeThread+0x5ab)
    /opt/mqm/lib64/libmqmcs_r.so(+0x16e93e)
    /usr/lib64/libpthread.so.0(+0x7e25)
    /usr/lib64/libc.so.6(clone+0x6d)
    
    MQM Function Stack
    zstAsyncConsumeThread
    zstAsyncConsume
    zstUserCallback
    zstConsumerFunction
    rstConsumer
    rstSendAsyncMessage
    rriCallSndExit
    xcsFFST
    
    ...
    
    Probe Id: RM052101
    Component: rriCallSndExit
    Probe Type: HALT6109
    
    On Windows, the amqrmppa process might crash with no FDC or with
    FDC XC130031. The stack trace extracted from the process dump
    file is likely to show the following call stack.
    
    00 03f7f774 100f7d04 amqrdlla!rriCallSndExit+0xc6b
    01 03f7f808 100f623a amqrdlla!rstSendAsyncMessage+0x894
    02 03f7f888 00274447 amqrdlla!rstConsumer+0xc0a
    03 03f7fab8 0027740e amqzi!zstUserCallback+0x537
    04 03f7fe3c 00000000 amqzi!zstAsyncConsume+0x232e
    

Local fix

  • Making one of the following changes and then restarting the
    affected client applications
    likely to workaround the problem.
    
    * Disable Channel statistics
    * Disable exit
    * Set SHARECNV to 0
    * If the client application is not using asynchronous consume
    feature set SHARECNV to 1
    

Problem summary

  • ****************************************************************
    USERS AFFECTED:
    * Users using a SVRCONN channel with an exit, the asynchronous
    consume feature and channel statistics enabled. If any one of
    these 3 elements is removed from the channel configuration or
    scenario, the issue is not observed.
    
    * Users using SHARECNV > 0 with SVRCONN channel exit and channel
    statistics enabled in MQ 7.x
    
    * Users using SHARECNV > 1 with SVRCONN channel exit and channel
    statistics enabled in MQ 8.0 and higher MQ versions
    
    * Users using SHARECNV = 1 with clients using async consume
    feature and SVRCONN channel exit and channel statistics enabled
    in MQ 8.0 and higher MQ versions might be affected by this
    problem.
    
    * Users using SHARECNV = 0 are not affected by this problem
    
    
    Platforms affected:
    MultiPlatform
    
    ****************************************************************
    PROBLEM DESCRIPTION:
    A design limitation of the SVRCONN channel implementation prior
    to MQ 9.1.2 meant that a potentially unsafe access was made to
    the channel status table in the scenario described above.
    

Problem conclusion

  • Design changes introduced in IBM MQ 9.1.2 have altered the
    implementation of SVRCONN channels such that this issue can no
    longer be encountered in this scenario.
    

Temporary fix

Comments

APAR Information

  • APAR number

    IT22101

  • Reported component name

    WMQ BASE MULTIP

  • Reported component ID

    5724H7241

  • Reported release

    750

  • Status

    CLOSED UR1

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2017-08-22

  • Closed date

    2019-07-11

  • Last modified date

    2019-08-29

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

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

Fix information

  • Fixed component name

    WMQ BASE MULTIP

  • Fixed component ID

    5724H7241

Applicable component levels

[{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSDEZSF","label":"IBM WebSphere MQ Managed File Transfer for z\/OS"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"7.5","Edition":"","Line of Business":{"code":"LOB45","label":"Automation"}}]

Document Information

Modified date:
31 March 2023