PI44960: CQS IS HUNG AFTER EMHQ STRUCTURE RECOVERY FAILS ON NON-MASTER WITH MESSAGE CQS0242E RC=43000080 15/07/23 PTF PECHANGE

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• An EMHQ structure recovery that is initiated with a CFRM policy
  EMHQ definition that is too small results in the master
  incorrectly reporting the structure recovery worked, and
  the non-master reporting that the recovery failed with
  message CQS0242E RC=43000080.
  

Local fix

• VERIFY THE CFRM POLICY DEFINITION FOR THE EMHQ STRUCTURE
  CORRECTLY REFLECTS THE CURRENT STRUCTURE SPECIFICATIONS.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All IMS and CQS V10 users of shared EMHQ.    *
  ****************************************************************
  * PROBLEM DESCRIPTION: After PK64986, An EMHQ structure        *
  *                      recovery that is initiated with a CFRM  *
  *                      policy EMHQ definition that is too      *
  *                      small results in the master incorrectly *
  *                      reporting the structure recovery        *
  *                      worked, and the non-master reporting    *
  *                      that the recovery failed with message   *
  *                      CQS0242E RC=43000080 size or ratio      *
  *                      mismatch.  After this, the non-master   *
  *                      CQS appears hung and processes no CQS   *
  *                      requests.                               *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  Two IMSs are cold started and their CQSs started after the
  coupling facility is reassigned back to the original CPU.  The
  EMHQ structure definition in the CFRM policy is smaller than it
  was when client data was previously checkpointed to the SRDS
  data set.
  
  The first CQS initializes and determines that the EMHQ structure
  needs to be rebuilt, becoming the structure recovery master.
  The structure recovery master detects a structure size/ratio
  mismatch with the data in the SRDS data set, and aborts the EMHQ
  structure recovery by sending an XCF message to the second CQS
  (non-master).  The master CQS erroneously continues performing
  the EMHQ structure recovery without aborting the rebuild for
  itself, and incorrectly reports that the rebuild completed
  successfully.
  
  The non-master CQS aborts the EMHQ structure recovery and issues
  message CQS0242E RC=43000080, followed by message CQS0244E.  Any
  subsequent structure process (such as structure checkpoint)
  fails because of the EMHQ structure recovery failure on the non-
  master CQS.  CQS appears to be hung and not able to do work,
  because the EMHQ structure recovery did not complete
  successfully.
  
  Depending upon the timing of CQS initialization for both CQSs,
  CQSXCF10 might fail with an ABEND0C4 for the non-master CQS.
  
  The CQS EMHQ structure recovery master should have aborted the
  EMHQ structure recovery for itself, after it detected the
  structure size mismatch error.
  
  Both of the CQSs should have attempted to initiate another EMHQ
  structure recovery, in order to correct the structure size
  mismatch.
  

Problem conclusion

• GEN:
  KEYWORDS:
   SYSPLEXSQ
  
  *** END IMS KEYWORDS ***
  CQSSTR00 is changed in subroutine RBLDABRT, just before it tests
  for a WAITRBLD=NO structure.  Code is added to first check if
  the structure is empty, and if it is, branch to process the XES
  rebuild abort, because all rebuild phases for a rebuild to an
  empty structure are processed within the XES rebuild, even for
  WAITRBLD=NO.
  
  
  CQSXCF10 is changed after label MSG_0240 to skip processing the
  rebuild XCF message, if this CQS is not aware of a rebuild in
  progress.  This can happen if CQS is initializing and didn't get
  previous rebuild messages.
  

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:

  PI45145 PI45146 PI45148 PI45149

Modules/Macros

• CQSSTR00 CQSXCF10
  

Fix information

• Fixed component name

  IMS V10

• Fixed component ID

  5635A0100

Applicable component levels

• R010 PSY UI29839

     UP15/08/08 P F508 «

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.