PI70508: CICS TASKS STUCK RUNNING WHILE DOING A NAMED COUNTER COMMAND. CAN NOT SHUTDOWN OR CANCEL CICS. CONTENTION FOR SYSZAXMC .

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• A dump of CICS shows that the CICS tasks that are stuck Running
  while doing a Named Counter command.  The L8 TCB on which the
  CICS tasks are running is waiting for an ENQ on resource
  SYSZAXMC .
  .
  The TCB that owns the SYSZAXMC resource is stuck in a PC 30D
  Wait out of module DFHNCMN near label XMBEGPUR .  This TCB is
  going through normal termination.  Module DFHNCMN has gotten
  control through a RESMGR exit driven at end of task.
  .
  In SUMM FORMAT, this TCB's non-dispatchability indicators look
  like this:
  .
  Top RB is in a wait
  SVC Dump is executing for another task
  Task is terminating and its mainline processing will not run
  again ( TCBDYING )
  Task is terminating and its mainline processing will not run
  again ( TCBENDNG )
  .
  The TCB has 1 RTM2WA and its summary is:
                                RTM2WA SUMMARY
                                --------------
  +001C  Completion code                    08000000
  +008C  Abending program name/SVRB address 00000000 00000000
  +0094  Abending program addr              00000000
  .
  This is all normal for a TCB that is going through termination.
  .
  The initial Linkage Stack entry is for a PC out of AXMSC.  If
  you take the R13 saved in that Linkage Stack entry.  Go there
  and follow the pointer at +4 to the prior save area.  In that
  save area, the saved R14 points to IGC0101C+01143A .  This is
  normal for when a TCB is within an End Of Task RESMGR exit.
  .
  The reason why a Cancel does not cancel this address space is
  because the TCB is hung on the SVC 1 Wait from code that was
  driven as part of a RESMGR End Of Task exit.
  .
  The last Linkage Stack is the PC 30D Wait where PASN and SASN
  are the ASID for the Named Counter Server. The PSW in this
  Linkage Stack points into the Named Counter Server ASID just
  after a PC instruction (B218E000) and the prior eyecatcher is
  AXMXMBEG.  The Reg1 saved in this last Linkage Stack is
  7ED60C58.  This is the ECB being waited on.
  .
  Earlier, perhaps hours earlier, in the kernel error data of the
  CICS region, there is an abendAICA:
  .
  00000029 10:58:44 005B ABEND                ---/0999 DFHKETIX
  0000002A 10:58:44 005B TRAN_ABEND_PERCOLATE ---/AICA DFHSR1
  0000002B 10:58:44 005B TRAN_ABEND_PERCOLATE ---/AICA DFHEBF
  0000002C 10:58:44 005B PROGRAM_CHECK        0C4/AKEA -noheda-
  .
  If you go to the full Kernel Error Data entry for the abend999,
  the MVS PSW and Registers there point to where CICS interrupted
  to force the abendAICA.  That PSW points just after an SVC 1
  wait and the prior eyecatcher is AXMSCREQ .  The saved Reg1 is
  the ECB being waited on.  It is 7ED60C58 which is the same as
  the ECB being waited on by the TCB that owns the SYSZAXMC ENQ.
  .
  The command that may be isued to
  check for ENQ being held a long time would be
  D GRS RES=(SYSZAXMC).  You would be looking for any holder that
  had not released the enqueue.
  
  
  Additional Symptom(s) Search Keyword(s): KIXREVDWZ
  

Local fix

• Use a Force to bring down the CICS region
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: Several transactions accessing a        *
  *                      Named Counter Server go into a wait     *
  *                      and cannot be cancelled and the CICS    *
  *                      region itself has to be force           *
  *                      cancelled.                              *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A transaction running named counter server code consumes over
  1 second of CPU and is abended AICA. This particular call to
  the named counter server, originating from DFHEBF, is being
  processed asynchronously, so the requesting L8 TCB is placed
  into a momentary SVC 1 wait whilst the named counter server
  code is executed. The AICA interrupt occurs during this wait.
  Later, when CICS decides to terminate the L8 TCB, named counter
  server code gets control via an EOT RESMGR exit, and tries to
  disconnect the TCB from the server. The RESMGR exit issues
  another wait but this will never complete. As it holds enqueue
  SYSZAXMC then this prevents other tasks from connecting and
  disconnecting to the Named Counter Server.
  ADDITIONAL KEYWORDS: ICVR U0999 abendAICA
  

Problem conclusion

• DFHEBF has been changed to ensure no AICA timeouts can occur
  whilst in named counter server code.
  

Temporary fix

•             *********
              * HIPER *
              *********
  FIX AVAILABLE BY PTF ONLY
  

Comments

APAR Information

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

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

  PI70581 UI41894

Modules/Macros

• DFHEBF
  

Fix information

• Fixed component name

  CICS TS Z/OS V5

• Fixed component ID

  5655Y0400

Applicable component levels

• R900 PSY UI41894

     UP16/10/29 P F610 ¢

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.