PH39154: CHIN ABENDS WITH ABN= 402-00000000 BUT DOESN'T END (HANG) WHEN DISPATCHER PROCESS DIES OR IS KILLED

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• CHIN abends with ABN= 402-00000000,LOC=.........????????+EA468
  but doesn't end (hang) when Dispatcher Process dies or is
  killed.
  
  After the dispatcher was abended, recovery was driven for the
  dispatcher TCB. The nature of the abend meant that a flag was
  set in the SDWA indicating that retry is not allowed, which
  resulted in the recovery routine for the dispatcher TCB
  percolating. This percolation resulted in the dispatcher
  termination ECB being posted, which released the CHINIT
  supervisor task.
  
  The CHINIT supervisor task detected that a dispatcher had
  terminated, and initiated CHINIT shutdown processing as
  designed. As part of shutdown processing, the supervisor task
  called CSQXSPRT to release all of the supervisor acquired
  resources.
  
  This involves scheduling dispatcher termination for each of the
  dispatchers by POSTing a dispatcher ECB. This fails when
  posting the already terminated dispatcher, resulting in a
  402-00000000 abend for the supervisor task. The supervisor task
  recovery gets invoked and retries the processing once more, but
  this again results in a 402-00000000 abend. Since this was the
  second abend for the supervisor task, its recovery routine
  percolates before finishing the necessary cleanup required to
  allow for a clean shutdown.
  
  This results in the CHINIT jobstep task waiting for attached
  subtasks to complete, but since they were never awoken by the
  supervisor task, they will never finish.
  
  IBM MQ Development: z/OS Service team identified some flaws in
  the recovery processing for the dispatcher tasks. The problem
  could occur for small timing windows in some dispatcher failure
  scenarios.
  .
  Additional keywords and symptoms:
  ABEND402 ABENDS402 S402 S0402
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of IBM MQ for z/OS Version 9       *
  *                 Release 2 Modification 0 and Release 3       *
  *                 Modification 0.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: An abend in a dispatcher TCB can cause  *
  *                      the CHINIT to hang indefinitely while   *
  *                      shutting down.                          *
  ****************************************************************
  When the dispatcher TCB abends with RETRY=NO, an ESTAE recovery
  routine percolates after issuing message CSQX112E. This results
  in an ECB being posted which is waited on by the supervisor.
  After this ECB is posted, the supervisor decides to terminate
  the CHINIT.
  As part of this process, a cleanup task is scheduled to run on
  every dispatcher without checking if all are active. When the
  supervisor tries to schedule the task on the terminated
  dispatcher, it abends 402-00000000.
  In this situation, the jobstep task still sees attached tasks
  so is left waiting indefinitely for an ECB.
  

Problem conclusion

• Shutdown routines are only scheduled on dispatcher TCBs which
  are running. This means that when a dispatcher abends, the
  CHINIT shuts down correctly without hanging.
  

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:

  UI81039 UI82343

Modules/Macros

• CSQXADPC CSQXDISP CSQXSPRT
  

Fix information

• Fixed component name

  IBM MQ Z/OS V9

• Fixed component ID

  5655MQ900

Applicable component levels

• R200 PSY UI81039

     UP22/07/01 P F206

• R300 PSY UI82343

     UP22/09/17 P F209  

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.