PK74341: RESTART OF BMP ABENDS U0102 AND IMS ALLOCATED EXTRANEOUS SLDS.

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Customer reported that IMS allocated extraneous SLDS datasets
  and restart the BMP abend U0102 because no checkpoint data is
  found. Here is the sequence of events that run into the problem.
  1) BMP is started and issues checkpoints
  2) BMP abends
  3) IMS control region is shut down and started again
  4) The previously abended BMP is restarted the next day.
     (important,at least one date change and all OLDS should have
     wrapped)
  5) BMP is restarted with CKPTID=LAST
  6) IMS allocates the first SLDS recorded in the current Prilog
  Record, which means the one from the IMS restart (only this one)
  7) BMP abends with U0102 because checkpoint data isn't found
  ---------
  Additional symptoms:
  MSGIRA200E AUXILIARY STORAGE SHORTAGE can be received
    as auxiliary storage reaches the 70% threshold
    repeatedly. IRA200E
  .
  The BMPs do not inform the logger of termination, therefore
  the SLDSREAD data remains allocated and is not freed resulting
  in massive auxiliary storage usage by the IMS control region.
  .
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All IMS V10 users                            *
  ****************************************************************
  * PROBLEM DESCRIPTION: IMS allocates many extraneous SLDS      *
  *                      in response to XRST call                *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  The first BMP was terminated with a /STOP REGION ABDUMP
  command while reading log data from a SLDS because of either
  dynamic backout or an XRST call.  The log data already read was
  retained because abnormal BMP termination did not inform the
  logger that it was terminating.  The logger kept the data for
  possible later requests from the region.
  
  When the BMP is restarted again and issues an XRST call for the
  last checkpoint, A DBRC bug causes DBRC to pick the wrong log.
  DBRC needs to find the PRISLD record from the previous execution
  but it does not use the current PRILOG time and cannot find the
  desired record.  It reverts to the current PRILOG and tells
  logger to use the earliest SLDS from that record.
  
  Logger cannot find the requested record in the SLDS and the
  XRST call fails with U0102.
  

Problem conclusion

• AIDS: RIDS/SYS RIDS/LOG SYS LOG
   GEN:
  KEYWORDS:
  
  *** END IMS KEYWORDS ***
  Calls have been added to DFSSABN0 to inform logger when a
  dependant region terminates while holding data read from
  SLDS.
  
  DSPOLD10 has been changed use the current PRILOG time on the
  DSPRLOCT call to find the PRISLD record from the previous
  IMS execution.
  

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:

  PK78193 UK43474

Modules/Macros

• DFSSABN0 DSPOLD10
  

Fix information

• Fixed component name

  IMS V10

• Fixed component ID

  5635A0100

Applicable component levels

• R010 PSY UK43474

     UP09/01/30 P F901 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.