PH45838: ISSUE IN CONTROL REGION WITH HANDLING CONNECTION WITH ASYNC READ OUTSTANDING

APAR status

• Closed as program error.

Error description

• There is a problem in the code in the management of the Flag
  that indicates an Async Read is outstanding.  This could cause
  an abend0C4 by allowing a connection with an outstanding Async
  Read request to close and cleanup it's resources before the
  Async Read is completed.   The abend0C4 would occur when the
  Async Read drives its completed method on the cleaned up
  Connection.
  
  The control region joblog could show a GPF or abend0c4 in JIT
  compiled method:
  com/ibm/ws/tcp/channel/impl/ZAioTCPChannel.readCompleted
  
  with native traceback:
  ZAioTCPChannelCppUtilities::readComplete
  CF_TCP_Request::readComplete
  CF_TCP_Connection_Object_Manager::processAsyncReadComplete
  ACR_ExecutionThread::RemoveAndProcessWork
  ACR_ExecutionRoutine
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:  All users of IBM WebSphere Application      *
  *                  Server                                      *
  *                  V9.0 and V8.5                               *
  ****************************************************************
  * PROBLEM DESCRIPTION: WebSphere Application Server for z/OS   *
  *                      ZAioTCPChannel.readCompleted received   *
  *                      an ABENDS0C4                            *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  During the processing of a request received over a
  channel chain with the ZAioTCPConnLink and
  XMemProxyCRInboundConnLink an asynchronous read request
  received an ABEND0C4 on the readComplete method calling from
  native code up to java.   The connection objects, which include
  the two ConnLink objects mentioned previously, have been
  released.  The cause of the ABEND 0C4 is that these Connection
  objects were cleaned up and released while an outstanding
  asynchronous read was still pending completion.
  The state of asynchronous reads and writes are maintained in
  the java objects associated with the connection.  A path was
  found in the processing of a read request that would
  erroneously reset the outstanding request flag, thus allowing
  the connection to cleanup without taking the necessary actions
  required for an outstanding request.
  The following FFDC indicates the path that caused the
  erroneous resetting of the outstanding request flag:
  FFDC Exception:java.io.IOException SourceId:
  com.ibm.ws390.xmem.proxy.channel.
  XMemProxyCRInboundConnLink.destroyCommon ProbeId:423
  Reporter:com.ibm.ws390.xmem.proxy.
  channel.XMemProxyCRInboundConnLink
  java.io.IOException: ZAIO INTERNAL: only one read operation
  permitted at a time on a connection link:
  com.ibm.ws.tcp.channel.impl.ZAioTCPConnLink
  

Problem conclusion

• Code has been modified to not reset the read outstanding flag
  
  on an error path through the ZAioTCPReadRequestContextImpl.
  
  
  
  The fix for this APAR is targeted for inclusion in fix pack
  
  8.5.5.22 and 9.0.5.13. 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

  WEBSPHERE FOR Z

• Fixed component ID

  5655I3500

Applicable component levels