PM72293: DFHTS1310 DETECTED AT EMERGENCY RESTART

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• At CICSTS42 emergency restart, abend DFHTS1310 1310_abend_9
  is detected. The error is issued during DFHTSAM CHECK process
  because the existing byte map is different from the copied one.
  These are the events that lead to the problem:
  during the previous run, the CESD task created an ICE for
  REQID DFHCESD to restart itself from.
  When CICS was emergency restarted, the backwards read of the
  log would have rebuilt the TSQ data for queue DFHCESD, along
  with an ICE to reflect the IC_DATA recovered from the log.
  At the end of the backwards log reading, end_delivery is called.
  DFHTSRM needs to know if queues are to be kept or not,
  as some that have been recovered need to be deleted.
  It calls ICRC to DELIVER_IC_RECOVERY_DATA to see whether
  IC wants the queue data to be kept or delete ICUS data
  associated with an ICE for security purposes.
  ICRC calls ICUS to CREATE_ICE_USER_EXTENSION to recreate this.
  However, this call fails with DISASTER for
  INVALID_FLATTENED_ICUS .
  DFHTSRM do some final checking before exiting,
  and calls perform_end_delivery_checks.
  This drives a TSAM CHECK. This copies the existing bytemap, then
  clears the real bytemap and rebuilds it from all the TSQs. The
  two byte maps should be identical.
  In this case however, there is a TSQ for DFHCESD that was
  rebuilt from the log, but because of the error from
  ICRC DFHTSRM didn't get as far as
  calling tsqueue_delete_recovered to get rid of the TSQ
  (if IC has told it to) or tsqueue_recover_aux_space
  to update the real bytemap to say this queue data was to be kept
  (if IC had told TSRM to keep the queue)
  so we have the situation where the TSQ exists, and its storage
  is represented in the rebuilt bytemap, whereas its storage was
  not reflected in the original bytemap.
  This validly causes CICS to the type 9 1310.
  

Local fix

• Define a TSMODEL for queue prefixed with DFHCESD to be
  non-recoverable .
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users                               *
  ****************************************************************
  * PROBLEM DESCRIPTION: DFHTS1310 1310_abend_9 on CICS          *
  *                      emergency restart.                      *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  Interval control fails to pass all the necessary data relating
  to a START request when calling temporary storage to store the
  FROM data for the START. This means that only the ICE contents
  are copied into the TSICDATA associated with the queue. If the
  temporary storage queue or REQID is defined as recoverable, it
  will be logged during syncpoint, and the TSQ and TSICDATA data
  will be hardened to the system logstream. If CICS is then shut
  down abnormally or cancelled, the subsequent emergency restart
  will drive DFHTSRM when the temporary storage log data is read
  back from the log. This will rebuild the queue and reconstruct
  the TSICDATA associated with it. DFHTSRM then calls DFHICRC to
  reconstruct the ICE and associated interval control data, from
  the TSICDATA passed to it from DFHTSRM. Since part of the data
  is not available to it, DFHICUS returns INVALID_FLATTENED_ICUS
  and DFHICRC returns DISASTER to DFHTSRM. CICS performs a final
  consistency check on the state of the temporary storage domain
  at this stage, but since the reconstructed temporary storage
  queue has not been finalised, as a result of this error, it
  detects that the environment is inconsistent. CICS issues
  message DFHTS1310 and terminates.
  KEYWORDS: MSGDFHTS1310 TS1310 1310
  

Problem conclusion

• Interval control has been changed to pass the necessary data to
  temporary storage for a START request. Temporary storage has
  been changed to correctly store this data in the TSICDATA, log
  it, recover it on an emergency restart, and present it back to
  interval control to rebuild the START request environment
  properly.
    CICS should be cold started having applied this fix, to ensure
  log data from previous START requests is not present.
  

Temporary fix

•             *********
              * HIPER *
              *********
  FIX AVAILABLE BY PTF ONLY
  

Comments

APAR Information

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

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

  PM79578 UK90680

Modules/Macros

•    DFHAPRX  DFHICP   DFHISRS  DFHISXM  DFHMIRS
  DFHMRXM  DFHTSAM  DFHTSBR  DFHTSCL  DFHTSDM  DFHTSDUC DFHTSDUF
  DFHTSDUS DFHTSMB  DFHTSP   DFHTSPT  DFHTSPTT DFHTSQR  DFHTSRM
  DFHTSST  DFHXTP   DFH62XM
  

Fix information

• Fixed component name

  CICS TS Z/OS V4

• Fixed component ID

  5655S9700

Applicable component levels

• R700 PSY UK90680

     UP13/01/09 P F301

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.