IT23684: XC130031 access violation and JVM termination with multi-threaded Java application

APAR status

• Closed as program error.

Error description

• A multi-threaded MQ classes for Java application that
  connections to and disconnects from a queue manager repeatedly
  using the BINDINGs transport mode experiences an access
  violation within an MQ library that results in an FFDC and JVM
  termination.  The FFDC from the error contains information such
  as:
  
  Probe Id          :- XC130031
  Component         :- xehExceptionHandler
  
  Process Name      :- C:\Program
  Files\Java\jdk1.7.0_80\bin\java.exe
  UserApp           :- TRUE
  Last HQC          :- 0.0.0-0
  Last HSHMEMB      :- 0.0.0-0
  Major Errorcode   :- xecF_E_UNEXPECTED_SYSTEM_RC
  Comment1          :- Access Violation at address
  000000000000001C when  reading
  
  The Windows Debugger (WinDbg) shows-
  
  BUCKET_ID:
  APPLICATION_FAULT_NULL_CLASS_PTR_READ_INVALID_POINTER_READ_mqe+6
  e0a9
  FAILURE_EXCEPTION_CODE:  c0000005
  FAILURE_IMAGE_NAME:  mqe.dll
  .
  .
  APPLICATION_FAULT_NULL_CLASS_PTR_READ_INVALID_POINTER_READ
  FAILURE_PROBLEM_CLASS:  APPLICATION_FAULT
  FAILURE_SYMBOL_NAME:  mqe.dll!Unknown
  FAILURE_PROBLEM_CLASS:  NULL_CLASS_PTR_READ
  

Local fix

Problem summary

• ****************************************************************
  USERS AFFECTED:
  This issue affects users who have applications that repeatedly
  connect to and disconnect from queue managers, using the
  BINDINGs transport mode, on multiple threads in the same
  process.
  
  
  Platforms affected:
  Windows
  
  ****************************************************************
  PROBLEM DESCRIPTION:
  There was a timing window whereby multi-threaded applications
  that created and closed connections repeatedly resulted in one
  thread deallocating a structure just prior to when another
  thread attempted to use it.  This would have occurred if the
  disconnecting thread was the last thread in the process to
  maintain an active connection and access the shared process
  memory before the newly connecting thread registered itself.
  This caused a memory access violation.
  
  If a core file can be obtained and the failing thread stack
  examined, then an access violation from the MQ function
  xtrEstablishTraceStatus would be a good indicator of this APAR.
  

Problem conclusion

• Additional locking has been added to the IBM MQ product code
  where the failure was occurring to prevent accessing memory that
  has been deallocated.
  
  ---------------------------------------------------------------
  The fix is targeted for delivery in the following PTFs:
  
  Version    Maintenance Level
  v8.0       8.0.0.11
  v9.0 LTS   9.0.0.5
  
  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 is sysrouted FROM one or more of the following:

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

Fix information

• Fixed component name

  IBM MQ BASE MP

• Fixed component ID

  5724H7251

Applicable component levels