PK85570: LONG CQS RESTART AFTER ABNORMAL TERMINATION

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• When CQS is restarted after an abnormal termination, restart
  time can can become excessive (many minutes or hours) if the
  amount of CQS log data required for restart is large.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All IMS V11 users of IMS shared queues or    *
  *                 shared EMH.  All users of CQS.               *
  ****************************************************************
  * PROBLEM DESCRIPTION: Excessive CQS restart elapsed time when *
  *                      the number of log records processed for *
  *                      restart is large.                       *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  CQS restart after a previous CQS abend can take an excessive
  amount of time (many minutes, or even hours) when the number of
  CQS log records that CQS must read for restart processing is
  large.  In one test, CQS took 3.5 hours to process 4 million log
  records during restart.
  
  During restart, CQS reads the log beginning at the last system
  checkpoint.  CQS builds internal control blocks to keep track of
  inflight work based on the log records that are read.  These
  blocks are stored internally in hash tables.  The hash table
  management code inserts a new table at the end of its hash table
  chain.  The code currently scans the entire chain to locate the
  end, and this does not scale well when the chains get long.
  
  Circumvention:  If CQS restart is taking a long time to
  complete, consider reducing the value specified on the SYSCHKPT
  parameter in the CQS Local Structure Definition PROCLIB member.
  This parameter specifies the number of CQS log records CQS
  writes between system checkpoints.  By making this number
  smaller, you will increase the frequency of CQS system
  checkpoints, which will reduce the number of log records CQS
  will need to read during CQS restart.
  

Problem conclusion

• GEN:
  KEYWORDS:
   SYSPLEXSQ
  
  *** END IMS KEYWORDS ***
  CQS internal hash table processing is changed to keep a "last on
  chain" pointer for each hash table slot.  This allows new hash
  table elements to be added at the end of the chain without
  having to scan the entire chain to locate the end.
  

Temporary fix

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

Comments

APAR Information

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

  PK85568

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

  UK47164

Modules/Macros

• CQSTBLHT CQSTBL00
  

Fix information

• Fixed component name

  IMS V11

• Fixed component ID

  5635A0200

Applicable component levels

• R100 PSY UK47164

     UP09/06/06 P F906 Ž

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.