PI34247: MEMORY LEAK FOR SUBPOOL 0, KEY2 STORAGE WHICH HAVE BEEN USED TO CONTAIN CPOOL ELEMENTS FOR MQ MCBC CONTROL BLOCKS

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• In our case, WAS CR Listener port got an exception due to
  MQRC_GET_INHIBITED, then it goes into retry for browse messages
  on the queue. But memory consumption was seen at every retry for
  browsing error. Finaly Subpool 0, Key 2 storage was filled up
  and WAS CR region abnormaly terminated with ABENDC78 rsn04.
  .
  ADDITIONAL KEYWORDS:
  SP0 KEY2
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of WebSphere MQ Version 7          *
  *                 Release 0 Modification 1 and                 *
  *                 Release 1 Modification 0.                    *
  ****************************************************************
  * PROBLEM DESCRIPTION: WAS application server running an MDB   *
  *                      that is unable to get messages from a   *
  *                      queue can fail with insufficient        *
  *                      storage available in the controller     *
  *                      address space.                          *
  *                                                              *
  *                      WAS controller address space can        *
  *                      generate message IEW4000I and storage   *
  *                      related abends like SC78 and S078.      *
  *                                                              *
  *                      Other applications can be affected by   *
  *                      this problem when they use async        *
  *                      consume.                                *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  WAS application server running a MDB using JMS or MQ RAR V7 or
  later and the MDB is unable to get messages from the input queue
  (for example queue is Get Inhibited) can use all the private
  storage below the BAR in the WAS controller region.
  
  When the MDB fails and is terminated, it does not issue MQDISC
  to disconnect from the queue manager. WAS retries restarting the
  MDB listener port at regular intervals causing the allocation of
  cell pool storage that it is only deleted when MQDISC is
  called.
  
  The cause of this problem is a build-up of blocks of size
  x'AD00' with eyecatcher "Pool for MCBEC" that are not deleted
  if MQDISC is not called.
  
  Applications using async consume can be affected by this problem
  when threads terminate without using MQDISC.
  

Problem conclusion

• The cell pool storage is now deleted at the termination of MDB
  listener thread (or any application thread that used MQCB to
  register a callback function) avoiding the storage leak.
  010Y
  100Y
  CSQMEOTC
  

Temporary fix

• *********
  * HIPER *
  *********
  

Comments

APAR Information

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

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

  PI35190 UI25428 UI25429

Modules/Macros

• CSQMEOTC
  

Fix information

• Fixed component name

  WMQ Z/OS V7

• Fixed component ID

  5655R3600

Applicable component levels

• R010 PSY UI25428

     UP15/04/30 P F504 ¢

• R100 PSY UI25429

     UP15/04/30 P F504 ¢

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.