PM90060: DFHMQ0480 MQCC=2 MQRC=2195 ON EVERY REQUEST TO MQ AFTER DISCONNECTING AND THEN RECONNECTING TO MQ

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• The following series of events happens:
  1)Your CICS region is connected to MQ and tasks are succesfully
  making MQ calls.
  2)You successfully disconnect CICS from MQ.
  3)You successfully
  re-connect to MQ but now every request to MQ results in
  DFHMQ0480
  .
  This problem happens when the following happens while CICS is
  connected to MQ.
  1)A task makes a call to MQ and then that task ends.  The L8
  TCB used by that task is now free for other tasks to use.
  2)A new task allocates that same L8 TCB for linking to an
  OpenAPI program. This new task does not make any MQ calls, and
  this new task never ends.
  3)While that new task never ends, CICS disconnects from MQ and
  reconnects to MQ.
  .
  For MQ to completely be done with the prior connection, MQ
  needs for the TCBs that have Identified to MQ to end. But since
  that L8 TCB is allocated to a never-ending task, it doesn't
  end. This causes MQ to fail each request from this CICS region
  with MQRC MQRI_COORD_HAS_TERMINATED (8503). However this return
  code is overlayed with 2195 prior to CICS taking the DFHMQ0480
  dump. CICS should put the original MQ reason code in the
  message or at least save the original MQ reason code somewhere
  before taking the dump.
  
  
  
  
  
  
  Additional Symptom(s) Search Keyword(s): KIXREVDWZ
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: DFHMQ0480 with MQCC=2 and MQRC=2195     *
  *                      received on all requests to WMQ after   *
  *                      disconnecting then reconnecting to WMQ. *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  CICS is connected to IBM WebSphere MQ and an instance of CKTI
  (the CICS-WMQ trigger monitor or task initiator) is started.
  However the initiation queue specified is invalid. As a result
  CKTI issues messages DFHMQ0101 and DFHMQ0110, and then
  terminates.
  Later CICS is disconnected from WMQ, which attempts to detach
  from all TCBs which have performed WMQ work. The TCB which was
  earlier used to run the CKTI task is now being used by a long
  running task elsewhere in CICS. As a consequence the CICS-WMQ
  connection does not close.
  Any subsequent WMQ requests will fail with message DFHMQ0480.
  Keywords: msgDFHMQ0480 MQRC_UNEXPECTED_ERRORMQINI INITQ
            msgDFHMQ0101 MSGDFHMQ0110
  

Problem conclusion

• DFHMQTSK has been modified such that after a failed attempt to
  open the WMQ initiation queue, the TCB that the CKTI task is
  running under is freed on task termination. As a result the TCB
  cannot be used by another task and therefore the CICS-WMQ
  connection can close.
  

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:

  PM96275 UK97724

Modules/Macros

•    DFHMQTSK
  

Fix information

• Fixed component name

  CICS TS Z/OS V4

• Fixed component ID

  5655S9700

Applicable component levels

• R700 PSY UK97724

     UP13/09/26 P F309

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.