PM70921: MAS FAILS WORKLOAD REGISTRATION AFTER EMERGENCY RESTART

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• After a system logger issue, all CMASes and MASes on the lpar
  failed.  Automation restarted the CMAS and CICS MASes with an
  emergency restart.  This resulted in all but one MAS
  successfully reconnecting to the CMAS and registering with the
  workload, thus preventing work from being routed to this region.
  The MAS agent code had to be initialized manually via COLM in
  order for work to be routed to the target MAS.
  
  In the CMAS EYULOG we see the following messages for this MAS:
  
   EYUTI0009I Topology warm start for MAS1 initiated
   EYUCL0009I ESSS ICT entered for Detach from MAS1
   EYUTS0001I Topology Disconnect for MAS1 Initiated
   EYUTS0003I Topology Disconnect for MAS1 Complete
   EYUWM0425I Target region (MAS1) has been terminated for
              Workload (WLD)
   EYUWM0421I Routing region (MAS1) has been removed from Workload
              (WLD)
   EYUCL0008I ESSS ICT entered for Attach to MAS1
   EYUTS0001I Topology Connect for MAS1 Initiated
   EYUCL0012I Connection of CMAS1 to MAS1 complete
   EYUTS0003I Topology Connect for MAS1 Complete
  
  The CMAS auxtrace showed this exception occurring in CLET
  around the time the MAS is being DETACHed:
  
  TASK  MTD  PREV TRAN OBJ LEVL  PT-ID  DEBUG   UOW CMAS/USR ENVR
  xxxxx CLET XLOP LEEI COM EXCP.    17  EXCEPT CPSM CMAS1    CMAS
  
  where
    CMPI_XCPT_XCLC   EQU  17  Change to data cache failed
  
  Additional Keyword(s) and Symptom(s):
  KIXREVGJT
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICSPlex SM V4R1M0 and V4R2M0 Users      *
  ****************************************************************
  * PROBLEM DESCRIPTION:    After a system logger issue, all     *
  *                      CMASes and MASes on the lpar failed.    *
  *                      Your automation tool restarted the      *
  *                      CMAS and CICS MASes with an emergency   *
  *                      restart.  All MASes reconnected to the  *
  *                      CMAS successfully but one MAS failed    *
  *                      to register the workload, preventing    *
  *                      the MAS from taking part in the CPSM    *
  *                      workload management.  After stopping    *
  *                      and restarting the MAS agent using the  *
  *                      COSH and COLM transactions, you saw     *
  *                      that transactions were again routed to  *
  *                      the MAS.                                *
  ****************************************************************
  * RECOMMENDATION: After applying the PTF that resolves this    *
  *                 APAR, all CMASes must be recycled to pick    *
  *                 up the new code.  Note that CMASes do not    *
  *                 need to be brought down and restarted at     *
  *                 the same time.                               *
  ****************************************************************
     When a CMAS terminates while MASes which take part in a
  workload remain active, the state of those MASes is set to
  LOSTCON, causing the workload to enter a frozen state and
  preventing changes to the structure of the workload.  However
  the MAS agent in the MASes does not shut down or unregister
  with the workload so they can continue to route and receive
  transactions even though the CMAS to which they connect is
  not available.
     In this case, the CMAS restarted before one of the MASes
  had completed termination, and during warm start of the Top-
  ology component, saw that a MAS was active and set a flag in
  the CICS System Descriptor Block (CSDB) indicating that the
  MAS did not need to reregister with the workload.  After the
  Topology warm start completed, the MAS finished shutting down
  and the CMAS processed the disconnect normally.
     When the MAS was restarted and connected with the CMAS, be-
  cause the flag had been set in the CSDB stating that the MAS
  was active when the CMAS started, it was not processed as an
  initial connect, so it did not register with the workload, nor
  was the workload notified that the MAS had reconnected, causing
  the workload to remain in a frozen state.
  

Problem conclusion

•    Module EYU0TSSE (TSSE - Topology Complete MAS Termination)
  was modified to clear the flag stating that the MAS had been
  active when the CMAS restarted.  Modules EYU0WMAT (WMAT - WLM
  Process AOR Termination) and EYU0WMTT (WMTT - WLM Process TOR
  Termination) were modified to unset the LOSTCON state and un-
  freeze the workload if a CMAS is terminating an AOR or TOR to
  which contact had been lost.
  

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:

  UK82344 UK82345

Modules/Macros

•    EYU0TSSE EYU0WMAT EYU0WMTT
  

Fix information

• Fixed component name

  CICS TS Z/OS V4

• Fixed component ID

  5655S9700

Applicable component levels

• R60M PSY UK82344

     UP12/10/08 P F210

• R70M PSY UK82345

     UP12/10/08 P F210

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.