PM30411: DURING CQS RESTART, AFTER AN OVERFLOW STRUCTURE REBUILD FAILURE CQS0242E RC=43000080, EXPECTED SUBSEQUENT 11/01/17 PTF PECHANGE

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• During an online CQS, the overflow structure and checkpoint data
  sets are lost. A subsequent CQS restart attempts to rebuild the
  overflow structure and fails with CQS0242E RC=43000080 structure
  attributes mismatch. Expected subsequent rebuilds never happen.
  However, CQS restart completes to CQS Ready message.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All IMS V10 users of shared queues and CQS.  *
  ****************************************************************
  * PROBLEM DESCRIPTION: After PK64986, during a CQS restart, an *
  *                      attempt to rebuild the overflow         *
  *                      structure fails with CQS0242E message   *
  *                      and return code 43000080 'mismatched    *
  *                      attributes'.  For this type of rebuild  *
  *                      failure, CQS will retry a new rebuild.  *
  *                      However the expected new rebuild for    *
  *                      the overflow structure never happened.  *
  *                      CQS restart completes to the CQS ready  *
  *                      message but leaves the data in the      *
  *                      overflow structure unrecovered.         *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  1- The reported problem:
  An online CQS system goes into an overflow mode and terminates
  abnormally.  Prior to the attempt of restarting this CQS, the
  overflow structure and system checkpoint data sets are deleted.
  The delete action simulates a structure failure and intends to
  initiate an overflow structure rebuild from z/OS logs and
  structure recovery data sets (SRDS) during subsequent restart.
  
  CQS restart finds overflow structure empty and SRDS valid and
  initiates a structure recovery.  The recovery finds the rebuild
  structure's entry value in SRDS and the structure block
  mismatched.  The mismatched structure attribute values could
  prevent a successful structure rebuild because the newly rebuilt
  structure might not be large enough to hold data.  So the
  rebuild is aborted with return code 43000080 'mismatched
  attributes' and CQS restart needs to initiate a new rebuild
  with the correct structure attributes.
  
  CQS restart decides to initiate a new rebuild based on the
  rebuild failure reason code returned in the structure block of
  the structure being rebuilt in the previous rebuild attempt.
  CQS restart incorrectly reads the rebuild failure reason code
  from the structure block of the structure not being rebuilt
  which is the primary structure.  Therefore CQS restart finds no
  rebuild failure reason code and incorrectly decides not to
  initiate a new rebuild for the overflow structure.  CQS should
  have the read rebuild failure reason code from the overflow
  structure block instead.
  
  CQS restart completes to the CQS ready message but does not
  recover the overflow structure as expected.  Messages that had
  been on the overflow structure are lost.  Further, the queue
  names in overflow are no longer accessible or usable.
  
  2- Another problem is found during unit testing:
  During the new rebuild of the overflow structure after the first
  rebuild failure RC=43000080, CQS reads the rebuild failure
  reason code from the wrong structure block which is the primary
  structure.  CQS finds no rebuild failure reason code and
  incorrectly decides to validate the rebuild structure attributes
  again.  The validation subroutine still finds mismatched
  attributes and decides to update the CQSSTRATTRIBUTES entry with
  the valid values.  However CQS finds the entry already exists
  because the entry was written in the first rebuild.  The write
  to the entry fails and CQS aborts the rebuild with RC=43000031
  'error write to CQSSTRATTRIBUTES entry' and restart abends with
  U0014-00000394 'unable to recover overflow structure'.
  

Problem conclusion

• GEN:
  KEYWORDS: ABENDU0014 U0014 0014
   SYSPLEXDS
  
  *** END IMS KEYWORDS ***
  CQSIST30 - Structure initialization and restart.
  
  In the REBUILD subroutine, in the loop that CQS tries to
  initiate a new rebuild based on the rebuild failure reason code,
  prior to reading the rebuild failure reason code STRRBF, removed
  the instruction that loads the primary structure block address.
  So CQS can read the reason code from the overflow structure
  block as it is loaded when getting into the loop.
  
  Added the instruction that reloads the primary structure block
  when exiting the loop and when returning to the caller to
  preserve the current logic.
  
  CQSSTR00 - Structure recovery processor.
  
  In the rebuild phase 2 at label PH203700, when CQS determines to
  validate the rebuild structure attributes based on the rebuild
  failure reason code, the code loads the current structure block
  of the structure that being rebuilt before reading the rebuild
  failure reason code.
  
  After reading the rebuild failure reason code, restored the
  primary structure block to preserve the current logic.
  

Temporary fix

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

Comments

APAR Information

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

  PM30102

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

  UK65131

Modules/Macros

• CQSIST30 CQSSTR00
  

Fix information

• Fixed component name

  IMS V10

• Fixed component ID

  5635A0100

Applicable component levels

• R010 PSY UK65131

     UP11/02/25 P F102 «

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.