PH25965: DEFERRED_ABEND AKC3 CESF TRANSACTION CAUSED ABENDB78 IN DFHKERPC AND DFHAP0001 0C4/AKEA IN DFHABAB

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• You are running CICS TS 5.3 and receive fault analyzer
  message:
   IDI0002I Module DFHSIP, CSECT KESGM, offset X'6E8':
            CICS abend AKC3
  followed by CICS message and a dump:
   DFHAP0001  An abend (code 0C4/AKEA) has occurred
              at offset X'FFFF' in module DFHABAB.
  .
  The dump shows the failing transaction was CESF, and
  the following kernel error information :
  .
  NUM  ERR_TIME KE_NUM ERROR TYPE     ERR_CODE  MODULE    OFFSET
  ==== ======== ====== ==========     ========  ======    ======
  001  12:30:11  0027  DEFERRED_ABEND ---/AKC3  DFHKESGM  0006E8
  002  12:30:12  0027  ABEND          B78/AKEB  UNKNOWN   UNKNOWN
  003  12:30:12  0027  PROGRAM_CHECK  0C4/AKEA  DFHTRPX   00029E
  .
  CESF was purged, but the purged had to be deferred, causing the
  DEFERRED_ABEND entry above.
  The  B78 abend occurred in DFHKERPC, trying to freemain an area
  originally getmained by module DFHKESGM, that had a length of
  x'3AC bytes out of MVS subpool (SP) 131.
  Code in DFHKERPC passed an incorrect address on the
  'storage release' PC call and also passed SP 132, instead
  of 131.
  The 0C4 occurs when DFHABAB's recovery routine is called, and
  it passes a bad length on a trace call to DFHTRPX.
  This also caused the CICS internal trace table to get overlaid.
  Additional Symptom(s) Search Keyword(s): KIXREVACC
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: ABENDB78 while processing AKC3          *
  *                      transaction abend.                      *
  ****************************************************************
  A purge is issued against a CICS task. The purge is deferred
  until control is returned to a CICS module which can handle
  deferred abends ( DFHAPLI1 ). The task is processing an
  EXEC CICS SIGNOFF. The task gets suspended on the SMLOCK when
  DFHKESGM issues a getmain for a 24-bit stack segment required
  by DFHESN (the module which handles SIGNOFF commands).
  When the suspend ends, a deferred abend is scheduled against
  the task due to the earlier purge request. When storage manager
  returns to DFHKESGM (which is running under the DFHAPLI1
  stack), the deferred abend is triggered - driving DFHAPLI1's
  recovery routine.
  DFHAPLI1 starts the AKC3 abend. DFHABAB is called which in turn
  percolates the abend to DFHAPLI1. The percolate request drives
  DFHKERPC. DFHKERPC contains macro logic to cleanup kernel stack
  state when DFHKESGM is interrupted trying to obtain a 24-bit
  stack segment.
  This cleanup logic attempts to release storage acquired by
  DFHKESGM but it references the wrong address and subpool which
  causes the ABEND B78.
  Additional keywords : SB78 B78 ABENDSB78
  

Problem conclusion

• The macro logic used by DFHKERPC to cleanup after a DFHKESGM
  24-bit stack segment processing error has been corrected to use
  the correct subpool, storage address and length on the freemain.
  

Temporary fix

Comments

APAR Information

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

  PH01489

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

  UI70560

Modules/Macros

• DFHKEDS  DFHKEIN  DFHKERKE DFHKERPC DFHKERRI DFHKETA  DFHKETIX
  

Fix information

• Fixed component name

  CICS TS Z/OS V5

• Fixed component ID

  5655Y0400

Applicable component levels

• R900 PSY UI70560

     UP20/07/18 P F007

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.