PH44495: THREAD HANG IN TRANSACTIONIMPL.RESUMEASSOCIATION DUE TO RACE CONDITION WHEN A TRANSACTION IS USED ACROSS MULTIPLE THREADS

APAR status

• Closed as program error.

Error description

• Under heavy load when WS-AT transactions are involved with back
  end systems such as CICS, when backend systems go offline, The
  WAS-AT transactions may hang.  Javacores show the thread is hung
  with the callstack shown below:
  
  [2/27/22 15:56:38:393 EST] 0000003b ThreadMonitor W   WSVR0605W:
  Thread "WebContainer : 54" (000001a2) has been active for 614946
  milliseconds and may be hung.  There is/are 1 thread(s) in total
  in the server that may be hung.
  at java.lang.Object.wait(Native Method)
  at java.lang.Object.wait(Object.java:189)
  at com.ibm.ws.tx.jta.TransactionImpl.resumeAssociation(Transac
  tionImpl.java:2978)
  at com.ibm.ws.tx.jta.TransactionImpl.resumeAssociation(Transac
  tionImpl.java:2931)
  at com.ibm.ws.Transaction.wstx.WSATControlSet.resumeSuspendedT
  ransactionAssociation(WSATControlSet.java:374)
  at com.ibm.ws.wstx.handler.WSATGenerator.handleInbound(WSATGen
  erator.java:227)
  at com.ibm.ws.wstx.WSTXClientTCMImpl.handleInbound(WSTXClientT
  CMImpl.java:100)
  at com.ibm.ws.wstx.WSTXClientTCMImpl.cleanupContext(WSTXClient
  TCMImpl.java:81)
  at org.apache.axis2.util.ThreadContextMigratorUtil.performCont
  extCleanup(ThreadContextMigratorUtil.java:191)
  ...
  

Local fix

• restart application server
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:  All users of IBM WebSphere Application      *
  *                  Server                                      *
  ****************************************************************
  * PROBLEM DESCRIPTION: Threads can hang in the transaction     *
  *                      service processing responses in         *
  *                      complex scenarios involving request     *
  *                      timeouts.                               *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A transaction may be associated with multiple threads in the
  same application server, for example when processing EJB IIOP
  requests locally on the same application server.  Requests
  propagating transaction context may also be made that target
  off-server entities such as EJBs and Web Services (when WS-AT
  propagation is configured) hosted on a different application
  server.
  If a transaction has been propagated both across multiple
  threads in the same application server and on a request to an
  external entity, and if a request failure such as a timeout
  condition occurs, then an error in the transaction service
  component's policing of the threads can result in a permanent
  hang of one of the threads while processing the response.
  

Problem conclusion

• The transaction service was modified so that additional
  checking was performed to ensure that the initial thread
  associated with the transaction delayed attempting to rollback
  the transaction until after all outstanding threads associated
  with the transaction were completed.  In addition, the
  transaction service was modified to ensure that responses were
  processed in the reverse order that the requests were made.
  
  The fix for this APAR is targeted for inclusion in fix packs
  8.5.5.22 and 9.0.5.12. For more information, see 'Recommended
  Updates for WebSphere Application Server':
  https://www.ibm.com/support/pages/node/715553
  

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

  WEBS APP SERV N

• Fixed component ID

  5724H8800

Applicable component levels