PI39539: TITLE=CSQ0,ABN=5C6-00C51194,S=00000864,C=W9700.800.CFM-CSQERWI1,M=CSQGFRCV,LOC=CSQELPLM.CSQERWI1+00005D00 DURING CF RECOVERY

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• The customer is running WebSphere MQ for z/OS V8.0 in a 6 way
  QSG. They lost one of their CF's and were unable to recover and
  eventually IPL'ed all images in the plex. Then 5 of the 6 queue
  managers successfully initialized and all failed structures were
  recovered. The one QMGR received the following abend (these
  appear to happen twice on QMGR):
  .
  DUMP TITLE=CSQ0,ABN=5C6-00C51194,S=00000864,C=W9700.800.CFM
             -CSQERWI1,M=CSQGFRCV,LOC=CSQELPLM.CSQERWI1+00005D00
  .
  00C51194 means CSQE_unexpected_XES_Err94
  .
  EC PSW at Time of Error: 477C2000 A8CC22A8  (UI20637)
                           Csect CSQERWI1 20140902 + x'5C30' in
  CSQELPLM
  Reg2,IXL_RET_CODE  00000008
  Reg3,IXL_RSN_CODE  00000864
  .
  IXLRSNCODEBADBUFSIZE EQU X'00000864' The buffer specified on an
                           IXLLIST or IXLCACHE request is not
  large enough to contain the data being read. No data is
  returned.
  .
  QMGR came down with:
  IEF450I CSQ0MSTR CSQ0MSTR - ABEND=S6C6 U0000 REASON=00C94510
  .
         00C94510 - CSQI_ABEND_CFM_ERROR in DMC
  .
  MSTR log shows the following series of errors:
  .
   CSQE033E !MQZ0 CSQERWI1 Recovery phase 1 failed for
     structure CICS connection name CSQEMQGPCSQ002, RC=00000008
     reason=00000864
   IXL016I CONNECTOR CSQEMQGPCSQ002 TO STRUCTURE MQGPCSQ_ADMIN
     TERMINATING:
   JOB CSQ0MSTR ASID 0184 REQUESTED DISCONNECT REASON=FAILURE.
  .
   IEA794I SVC DUMP HAS CAPTURED:
   DUMPID=010 REQUESTED BY JOB (CSQ0MSTR)
   DUMP TITLE=CSQ0,ABN=5C6-00C51194,S=00000864,C=W9700.800.CFM
  -CSQERWI1,M=CSQGFRCV,LOC=CSQELPLM.CSQERWI1+00005D00
  .
  This is similiar (same abend/same module and offset) to APAR
  PI27192/UI22792 but the the IXL reason code is different.
  .
  The change team looked at the dumps and found that the abend in
  CSQERWI1 occurs because the admin structure contains an
  unexpectedly large entry for an eUOWD unit of work record.
  eUOWDs should be at most x1000 bytes long, but entries 2 and 3
  on List number 4 are x1100 bytes. This means that the request by
  CSQERWI1 to read these entries provides a buffer which is too
  small to hold them, and the read call is rejected with
  ixlRsnCodeBadBufSize.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of WebSphere MQ for z/OS Version 8 *
  *                 Release 0 Modification 0                     *
  ****************************************************************
  * PROBLEM DESCRIPTION: On queue manager initialisation there   *
  *                      may be an abend S5C6-00C51194 in        *
  *                      CSQERWI1 when trying to recover the     *
  *                      CSQ_ADMIN CF structure.                 *
  *                      Alternatively CSQIUOWA may abend with   *
  *                      0C4.                                    *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  CSQILCSR and CSQILPRC contain a double '-' in the calculation
  of eUOWDSize which is interpreted as an addition.
  This means that the eUOWD checkpoint record increases by twice
  the length of the LRH each time it is written to the LOG.
  Eventually this record can exceed the expected length of
  x'1000' and this can cause an 0C4 in CSQIUOWA, or a 00C51194
  in CSQERWI1 if it needs to recover the CSQ_ADMIN structure.
  

Problem conclusion

• CSQILCSR and CSQILPRC have been changed to correctly
  calculate the eUOWDSize.
  000Y
  CSQILCSR
  CSQILPRC
  

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:

  UI33675

Modules/Macros

• CSQILCSR CSQILPRC
  

Fix information

• Fixed component name

  WMQ Z/OS 8

• Fixed component ID

  5655W9700

Applicable component levels

• R000 PSY UI33675

     UP16/01/08 P F601 ¢

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.