PI73569: CPSM SYSTEM TRANSACTION TERMINATION IF RECOVERY ROUTINE UNABLE TO OBTAIN STORAGE DUE TO SOS

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Your CICS region is under stress and the CICS Short On Storage
  condition occurs. At that moment, one of the critical CPSM
  system tasks encounters an unexpected condition and goes
  through normal CPSM recovery.
  .
  In the recovery routine, an EXEC CICS GETMAIN is issued for a
  work area. This fails due to the SOS condition. This results in
  the CPSM agent code terminating in the region.
  .
  CPSM should protect its critical tasks and allow for this
  situation by preallocating an emergency workarea in case no
  other storage is available.
  .
  Additional Symptom(s) Search Keyword(s): KIXREVSVR
  

Local fix

• If this is a MAS, restart the MAS agent code manually by
  issuing transaction COLM, after you have addressed the cause of
  the SOS.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * All CICSPlex SM V5R1M0 Users                                 *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * If an abend occurs in a MAS when the MAS is under stress due *
  * to a CICS Short On Storage (SOS) condition, the MAS agent    *
  * may terminate with one of the following messages being       *
  * issued:                                                      *
  *                                                              *
  *   EYUXS1017E The controlling task for this region,           *
  *              transaction ( CONL ), is terminating abnormally *
  *                                                              *
  *   EYUXS1020E Terminating the MAS agent due to the            *
  *              termination  of the task for transaction        *
  *              ( COI0 )                                        *
  *                                                              *
  *   EYUXS1020E Terminating the MAS agent due to the            *
  *              termination  of the task for transaction        *
  *              ( COIE )                                        *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * After applying the PTF that resolves this APAR, all MASes    *
  * must be restarted.  Note that the restarts do not need to    *
  * occur at the same time.                                      *
  ****************************************************************
  When an abend occurs in a MAS, module EYU9XLRV is called as part
  of CICS HANDLE ABEND processing. XLRV requires a workarea to
  perform its processing, so issues an EXEC CICS GETMAIN for the
  storage. If the GETMAIN request fails, for example, if CICS is
  in an SOS condition, then XLRV cannot continue recovery for the
  task, and issues an EXEC CICS ABEND CANCEL to kill the task.
  This drives the CPSM task recovery TRUE, EYU9XSTR.
  
  XSTR will perform cleanup for the task, and will then determine
  if the abending task is the MAS long running task (CONL), the
  MAS communications task (COI0) or the MAS heartbeat task (COIE).
  If it is one of these three tasks, then the MAS cannot function
  without it, so XSTR requests termination of the MAS agent.
  

Problem conclusion

• To avoid MAS agent termination in these cases, changes have been
  made to pre-allocate the area required by XLRV for the CONL,
  COI0 and heartbeat tasks at MAS initialization.
  
  -  new copybook EYUBXLMR has been created to map the three
     pre-allocated recovery areas.
  
  -  module EYU9XLBE (XLBE), which is called under the CONL task
     to perform some of the processing for MAS initialization,
     will allocate the area mapped by the new EYUBXLMR copybook,
     and store the address of the area into the CPSM Kernel
     Linkage anchor block (XLWA).
  
     XLBE will also turn on a flag in the CPSM object process
     block (OPB) for the CONL task to indicate it is the CONL
     task.
  
  -  module EYU9XLOP (XLOP), which is the starting program for the
     COI0 and heartbeat tasks, will turn on a flag in the OPB for
     the task to indicate that the task is the COI0 or heartbeat
     task.
  
  -  module EYU9XLRV (XLRV), which performs CICS HANDLE ABEND
     recovery for CPSM tasks, will check the OPB for the abending
     task, and if it is the CONL, COI0 or heartbeat task, will
     retrieve the address of the EYUBXLMR from the XLWA, and set
     the recovery address for the task from this area, instead of
     issuing a CICS GETMAIN to acquire the storage.  When recovery
     is complete, it will bypass the CICS FREEMAIN of the area.
  
  -  module EYU9XLEV (XLEV), which is the last CPSM module that
     processes at MAS agent termination, has been updated to free
     the EYUBXLMR just prior to returning to CICS.
  

Temporary fix

Comments

APAR Information

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

  PI56812

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

  UI44796

Modules/Macros

• CJA9XLEV CJA9XLRV CJB9XLEV CJB9XLRV CJC9XLEV CJC9XLRV CJD9XLEV
  CJD9XLRV EYU9XLBE EYU9XLEV EYU9XLOP EYU9XLRV
  

Fix information

• Fixed component name

  CICS TS Z/OS V5

• Fixed component ID

  5655Y0400

Applicable component levels

• R80M PSY UI44796

     UP17/02/21 P F702

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.