PM22555: DFHSM0002 (CODE X'030F') QUICKCELL_FREEMAIN_QPF_ALREADY_FREE CRSQ ICE

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• You received DFHSM0002 A SEVERE ERROR (CODE X'030F')
  abend shortly after control was given to CICS(startup was warm).
  Transaction running was CRSQ.
  You see the following trace exception entry:
   SM 030F SMGF  *EXC* - Quickcell_freemain_QPF_already_free
           FUNCTION(FREEMAIN)
           SUBPOOL_TOKEN(16F90890 , 000000A8)
           ADDRESS(183C6618) REMARK(ICE)
  
  It was trying to free an ICE that was already freed.
  The problem is that the code run by CSSY is interrupted.
  When task CSSY enters DFHICP for expiry anaysis, it takes a copy
  of the CSAICEBA and places it into ICECBAR so it can
  manipulate it later for running down the chain. ICECBAR is
  updated during processing to be the value of the ICE being
  processed. However because task CSSY it is interrupted before
  it can make the attach, DFHICP is unaware that the copy of
  the ICE contained in ICECBAR is no longer valid as it is no
  longer on the ICE chain.
  The events needed for this to occur is for both CRSQ tasks
  that are on the ICE chain to timeout so close together,
  that the second CRSQ task times out and is driven before the
  first CRSQ task has begun to run and had a chance to clean
  up the ICE chain.
  Additional Symptom(s) Search Keyword(s): KIXREVPAA
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: During a WARM restart, CICS encounters  *
  *                      DFHSM0002 A severe error (code X'030F'  *
  *                      or code X'0305') has occurred in module *
  *                      DFHSMGF.                                *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  CICS is started using a WARM startup. Recovery Manager reads
  back the log to recover resources, including any ICEs from the
  previous run of CICS. An ICE is recovered for CRSQ, which
  expires almost immediately (less than one second to go). The
  initialisation task creates a fresh ICE for CRSQ, and places it
  on the chain. The first ICE (from the log) expires. DFHICP
  attaches a CRSQ task using this ICE, then takes a copy of the
  new head of the ICE chain. CRSQ runs, unchaining and freeing any
  other CRSQ ICEs it finds on the ICE chain. After this completes,
  DFHICP attaches another CRSQ task, for the ICE from the
  initialisation task. It does this because it took a copy of the
  head of the ICE chain earlier on, not realising that this ICE
  has already been freed by the first CRSQ task. When the second
  CRSQ task completes, it attempts to free the ICE controlling it,
  leading to the invalid freemain error.
  
  Additional Keywords: msgDFHSM0002 SM0002 030F 0305
  

Problem conclusion

• DFHICRC has been amended so that ICEs for the CRSQ task are not
  recovered from the log during a WARM restart.
  

Temporary fix

• FIX AVAILABLE BY PTF ONLY
  

Comments

APAR Information

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

  PM20644

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

  UK60654

Modules/Macros

•    DESICRC  DFHICRC
  

Fix information

• Fixed component name

  CICS TS Z/OS V4

• Fixed component ID

  5655S9700

Applicable component levels

• R600 PSY UK60654

     UP10/10/01 P F009

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.