PH40878: INCONSISTENT BEHAVIOR OF SOME QUEUE MANAGERS IN A QSG, DURING CF REBUILD ATTEMPT, CAUSING ABN=026-08110102

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Delay of commit processing of 1 Queue Manager belonging to a
  QSG, due to a Log Full condition ( CSQJ111A ), prevents other
  Queue Managers from processing failure events, due to a CF
  structure failure.
  This prevents an expected CF rebuild, until the 2-phase commit
  is complete.
  If the WAIT exceeds a pre-defined XCF Policy, XCF terminates
  the hung process, and is working 'as-designed'.
  However, in this particular case, the termination of
  additional queue managers that had not lost connectivity to the
  structure, nor filled their logs, is not expected.
  This 'inconsistent' behavior needs to be handled in a more
  elegant way.
  

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: Several queue managers are terminated   *
  *                      S026-08110102 following a partial loss  *
  *                      of connectivity to a subset of the      *
  *                      queue managers.                         *
  ****************************************************************
  A queue manager experienced a partial loss of connectivity to
  an application structure, and prepared to disconnect from the
  structure. However processing of the EEPLLossConn event was
  delayed due to the logs having filled while a commit was in
  progress.
  While this processing was delayed another qmgr disconnected from
  the structure due to the same loss of connectivity, and all
  queue managers that remained connected (including the delayed
  queue manager) received a DiscFailConn event.
  
  One of them started Peer Level Recovery (PLR) phase 1, and
  waited for all connected queue managers to acknowledge the
  associated USync event.
  This resulted in PLR phase 2 being delayed due to the original
  queue manager's full log condition, and consequently can result
  in other queue managers participating in the PLR operation, in
  addition to the one that was hung, being terminated
  abnormally due to the S026-08110102 abends.
  

Problem conclusion

• Queue managers processing a loss of connectivity event will
  not participate in new PLR operations for an affected structure,
  allowing other participating queue managers to complete PLR
  processing and respond to the DiscFailConn event.
  

Temporary fix

Comments

• ×**** PE23/03/23 FIX IN ERROR. SEE APAR PH53483  FOR DESCRIPTION
  

APAR Information

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

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

  UI79255 UI81566

Modules/Macros

• CSQE197M CSQESEX  CSQEWRKT
  

Fix information

• Fixed component name

  IBM MQ Z/OS V9

• Fixed component ID

  5655MQ900

Applicable component levels

• R200 PSY UI79255

     UP22/04/01 P F203

• R300 PSY UI81566

     UP22/08/03 P F208 ¢

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.