PI19230: WMQ 7.1: WHEN STARTING A QMGR AFTER THE CICS REGION THE CICS WMQ CONNECTION WILL NOT BE RE-ESTABLISHED

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• When customer is starting MQ after the the CICS region (or
  bounce WMQ while CICS is up) the CICS WMQ connection will not
  be re-established:
  
  DFHMQ0341 STOP requested by alter monitor CKAM
  DFHMQ0107 A request to end CKTI has been received. CKTI ended.
  DFHMQ0107 A request to end CKTI has been received. CKTI ended.
  DFHMQ0334 Adapter shutdown successful.
  DFHMQ0342 CONNECT received from alert monitor.
  DFHMQ0308 MQNAME ABC is stopped. Connect request deferred.
  DFHMQ0380 No active connection. STARTCKTI rejected.
  DFHMQ0380 No active connection. STARTCKTI rejected.
  
  This problem occurs when a queue-sharing group name with less
  than 4 characters is used (for example 'ABC').
  When the queue-manager ABC1 is stopped CICS issues another
  identify request for ABC, passing the address of an ECB to
  be posted when a queue-manager in group ABC becomes avaiable.
  This ECB is chained off a GATE control block for the group,
  which has the group name stored as 'ABC'.
  However, when a queue-manager in group ABC does start, it is
  looking for a GATE control block for 'ABC§' (internally the
  queue sharing group name is passed to 4 characters
  with at-signs). This means that the ECB
  provided by CICS is not posted, and hence it does not
  automatically reconnect.
  

Local fix

• Manual activation of the CICS MQ connection (e.g. using CEMT)
  does work:
  DFHMQ0307 Successful connection to queue manager ABC1
  release 0710 group ABC
  
  Also when setting the MQCONN to RESYNCMEMBER(YES) does work.
  So in our production environment GROUPRESYNC does not work and
  must be backed out.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of WebSphere MQ for z/OS Version 7 *
  *                 Release 1 Modification 0.                    *
  ****************************************************************
  * PROBLEM DESCRIPTION: CICS does not reconnect to QSQ with     *
  *                      MQCONN having Resyncmember set to       *
  *                      Groupresync, and the QSG name less than *
  *                      4 characters long.                      *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  In CICS a MQCONN has been defined for a QSG where the QSG name
  is less than 4 characters in length. and with Resyncmember set
  to Groupresync.
  A QMGR that it is connected to in the QSG is stopped and that
  was the only active QMGR in that QSG. When that QMGR, or any
  other QMGR in the QSG is started, CICS does not automatically
  reconnect to the QSG. It can be reconnected by doing a CEMT SET
  MQCONN CONNECTED.
  The problem occurs because the GATE control block that has the
  connection information holds the QSG name in an internal format
  where the name is right padded with at signs.
  

Problem conclusion

• CSQ3IDGA has been altered to look for a GATE using a QSG name
  right padded with the at sign if the QSG name is less than four
  characters long.
  If it does not find any it will create a new GATE with a
  groupname including at signs.
  100Y
  CSQARIB
  CSQ3IDGA
  

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:

  PI27572

Modules/Macros

• CSQARIB  CSQ3IDGA
  

Fix information

• Fixed component name

  WMQ Z/OS V7

• Fixed component ID

  5655R3600

Applicable component levels

• R100 PSY UI20766

     UP14/10/07 P F410

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.