PI53066: QMGRS IN A PARTIAL REPOSITORY HUNG DURING SHUTDOWN.

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• The customer attempted to stop all queue managers
  in a sysplex. Only one qmgr with the full repository
  came down gracefully.
  It seems the other qmgrs hung (or stalled) during
  shutdown. Eventually the customer had to use the
  force command to bring them down.
  .
  When the STOP QMGR MODE(FORCE) command is issued,
  the queue-manager tries to put an event message to
  SYSTEM.ADMIN.COMMAND.EVENT. However, as there
  is no local queue of this name, a query is sent to
  the full repository and CSQMZLOO goes into a 1 second
  wait.
  However, as a STOP MODE(FORCE) has been issued CSQMZLOO
  is never woken from the wait.
  This results in the shutdown hang.
  .
  Additional Symptom(s) Search Keyword(s):
  

Local fix

• No Local Fix
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of WebSphere MQ for z/OS Version 8 *
  *                 Release 0 Modification 0.                    *
  ****************************************************************
  * PROBLEM DESCRIPTION: Queue managers that are partial         *
  *                      respositories in a cluster, with        *
  *                      command events turned on, but without a *
  *                      local SYSTEM.ADMIN.COMMAND.EVENT queue  *
  *                      for the event messages to be put to,    *
  *                      may hang during shutdown, if STOP       *
  *                      MODE(FORCE) is used without a prior     *
  *                      stop MODE(QUIESCE).                     *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  During shutdown with MODE(FORCE), a command event message will
  be put the command event queue (SYSTEM.ADMIN.COMMAND.EVENT) if
  event messages are turned on for commands. However in the case
  of where the queue is not local, and the queue manager is a
  member of cluster, CSQMZLOO is called to look up the queue
  within the cluster. During the lookup, if no existing
  subscription exists, a sleep may be entered. However this isn't
  cancelled by the force command, resulting in the queue manager
  hanging during shutdown.
  

Problem conclusion

• The sleep call within CSQMZLOO has been changed, to be allowed
  to be terminated in the case of a MODE(FORCE) stop of the queue
  manager.
  000Y
  CSQMZLOO
  

Temporary fix

Comments

APAR Information

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

  PI48946

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

  UI34088

Modules/Macros

• CSQMZLOO
  

Fix information

• Fixed component name

  WMQ Z/OS 8

• Fixed component ID

  5655W9700

Applicable component levels

• R000 PSY UI34088

     UP16/02/03 P F602

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.