PK10822: DFHDP0001 AZI6 AT OFFSET X'0340' IN MODULE DFHDPLM FOLLOWED BY AN S0C4 WITH ACTIVE LOAD MODULE=DFHSIP

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• The CICS shutdown process can abend with a DP0001, followed by
  an abnormal termination of the CICS region. This may happen if
  the IBM Debug Tool profile dataset DFHDPFMB is not local.  If
  the region where the DFHDPFMB is located is also being
  terminated, or if the region shutting down is for some reason
  unable to get over to the region where DFHDPFMB lives, then this
  problem may occur.
  Additional Symptoms: MSGDFHPD0001 ABENDAZI6 CRASH TERMINATE
  Local Fix: Either define DFHDPFMB locally or ensure that the
  region which owns the DFHDPFMB file is shut down after any other
  region which needs access to this file.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All                                          *
  ****************************************************************
  * PROBLEM DESCRIPTION: CICS abends while CIDP is being used    *
  *                      to attempt to inactivate debug          *
  *                      profiles.                               *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  CIDP was used to attempt to inactivate debug profiles. This
  drove DFHDPIN, which in turn called the DP domain module DFHDPLM
  to browse the dataset. The profile dataset DFHDPFMB was defined
  as a remote file, and so CICS attempted to function ship a
  startbr operation to it. During the IRWAIT wait on the MRO link,
  the task was purged. DFHZIS2 issued an AZI9 abend. This drove
  the recovery routine for DFHFCRF, then for DFHFCFR. It then
  reached DFHDPLM's recovery routine. This issued a dfhkern
  recovery_exit (not a percolate). As such, control did not
  percolate back to the APLIREC recovery routine. This meant that
  TCALCDSA was not restored to the AP domain stack pointer that
  DFHDPIN was using. The abend in progress flag TCAABREC was also
  left on. DFHDPIN tested the response to the DFHDPLM call, and
  the otherwise clause was reached, which drove an 0C4. CICS then
  terminated as part of this subsequent abend.
    Note - the same problem could occur for a readnext too.
  An additional symptom is:
  An FOR suffered a loop condition after a CSMI transaction
  received abend AZI6 because of the connected region was
  shutting down. The STACK for DFHXMIQ is pointing to
  itself causing the CSMI task to loop.
  
  KEYWORDS: s0C4 abends0c4 abend0c4 dfhdpcm tczabrec tczlcdsa
  MSGDFHDP0001 DFHDP0001 DP0001 DP 0001 AZI6 AZI
  

Problem conclusion

• DFHFCFR has been changed to ensure tcalcdsa is restored
  upon exit from its recovery routine.
    DFHDPCM has been changed to return the dplm_file_error
  exception from the recovery routine used by DFHDPLM.
    DFHDPIN has been changed to set off the abend in
  progress indicator if this exception is returned.
  

Temporary fix

• 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:

  UK09540 PK38649

Modules/Macros

•    DESDPCM  DESDPIN  DESDPLM  DFHDPIN  DFHDPLM
  DFHFCFR
  

Fix information

• Fixed component name

  CICSTS 3.1 Z/OS

• Fixed component ID

  5655M1500

Applicable component levels

• R400 PSY UK09540

     UP05/12/02 P F512

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.