PK48766: STARTING XE AGENT CAUSES 878-10 GETMAIN ERR'S

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Shortly after starting the xe agent task the omegamon task gets
  CI0970: OMEGAMON SUBTASK ABEND CODE=84878000
  PSW=070C100081396A36
  CI0971: PROGRAM NAME=KO2CCAF2
  IEA705I ERROR DURING GETMAIN SYS CODE = 878-10 OMIIO2 OMIIO2 00
  
  In the OMPE server joblog will be a high volume of these
  messages
  KO2O1918I DPX4+ IFI COLLECTOR (CAF) CONNECTION ESTABLISHED
  KO2O1922I DPX4+ IFI COLLECTOR (CAF) FORCIBLY DETACHED
  
  IEA705I ERROR DURING GETMAIN SYS CODE = 878-10 KD2RTC KD2RTC 00
  IEA705I ERROR DURING GETMAIN SYS CODE = 80A-10 KD2RTC KD2RTC 00
  
  A large amount of subpool 0 (SP0) key 7 storage will contain the
  eyecatcher I#ISWA24.
  
  Private storage below the line will be exhausted
  
  A review of the dump may show additional abends in DSNACAF or
  other DB2 module involved in call attach or in IDA0192C for DB2
  Media Services. This has not been shown to effect the DB2
  subsystem, just the OMPE servers attempt to collect trace data
  from DB2.
  
  There may be evidence of RACF failures within the OMPE server
  joblog, particularly if historical is attempting to switch logs
  KO2R0140W SWITCHING FROM DATA SET
  SYSS.CCC.AMXE.DIPH.DPX4.RKD2VS02             - DATASET IS FULL
           (SWITCT  )
  IRR401I 878 ABEND DURING RACF PROCESSING OF  984
    NEXT REQUEST FOR ENTRY SYSS.
  
  IRR401I 878 ABEND DURING RACF PROCESSING
  
  There may be evidence of DB2 Media Manager failures
  *IEA794I SVC DUMP HAS CAPTURED:  368
   DUMPID=022 REQUESTED BY JOB (KD2RTC  )
   DUMP TITLE=MEDIA MANAGER SERVICES(IDA0192C)ABEND S878, U0000, R
              EGISTER 15 = 00000010
  
  The storage overlay pattern may look like
  PSW=S  TCB=S
  at address 00C1C2E9
  
  
  
  The 878 abends are caused by storage below the line being
  exhausted. Prior to UK27357/UK28358 for this APAR, the OMPE
  server main TCB spawned child TCBs that shared the Subpool 0 Key
  7 storage. That meant it was the responsibility of those child
  TCB to free their storage at termination. If they did not free
  the storage, the storage remained allocated to the parent TCB.
  
  UK27537 changed this behavior so that the child TCB do not share
  the Subpool 0 Key 7 storage. Under this change if the child TCB
  terminates without freeing its own storage, VSM will do the
  cleanup and free the storage of the child task.
  
  The effect of these 878 abends can extend beyond the OMPE server
  address task. Service tasks such as those provided by OMNIMON
  base to post messages to the joblog or system dump services are
  also dependent on being able to get storage within the OMEGAMON
  DB2 address space. If storage is not available for the service
  tasks, the results can be unpredictable.
  
  00C1C2E0: 321C9540 F0134780 32E24040 D7E2E67E |..n 0....S  PSW=
  |
  00C1C2F0: E24040E3 C3C27EE2 40703224 95D7F012 | S  TCB=S...nP0.
  |
  
  S80A 80A  storage overlay
  
  Also see APAR OA23551
  

Local fix

• A work-around is to stop/restart the common collector:
  
  F taskname,P COMMCOLL
  F taskname,S COMMCOLL
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: OMEGAMON XE for DB2 PE/DB2PM                 *
  *                 users of component:                          *
  *                 -   OMEGAMON Server                          *
  *                 -   Common Collector                         *
  *                 -   OMEGAMON DB2 agent                       *
  ****************************************************************
  * PROBLEM DESCRIPTION: S878-10 abend followed by S80A-10       *
  *                      ABENDs in Omegamon Server after         *
  *                      starting agent.  There may be hundreds  *
  *                      of KO2O1918I and KO2O1922I messages     *
  *                      just prior to the abends:               *
  *                      KO2O1918I db2+ IFI COLLECTOR (CAF)      *
  *                      CONNECTION ESTABLISHED                  *
  *                      KO2O1922I db2+ IFI COLLECTOR (CAF)      *
  *                      FORCIBLY DETACHED                       *
  ****************************************************************
  * RECOMMENDATION: Apply this PTF.                              *
  ****************************************************************
  Below-the-line storage was being exhausted.
  

Problem conclusion

• Prior to this change, subtasks of the OMPE server shared private
  below the line storage in Subpool 0 Key 7 with the OMPE server
  main task. If the subtask terminated without freeing the
  storage, the storage remained allocated to the main task.
  
  After this PTF is applied the subtasks will be started with
  their own Subpool 0 Key 7 storage.   If the subtask does not
  free its storage at termination of the subtask, VSM will free
  the storage.
  
  KEYWORDS :
  ABENDS878 GETMAIN ABENDS80A-10 SERVER COMMCOLL COMMON COLLECTOR
  KO2O1918I KO2O1922I
  

Temporary fix

Comments

APAR Information

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

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

  UK27357 UK27358

Modules/Macros

•    KO2AINIB
  

Fix information

• Fixed component name

  OM XE DB2PE/PM

• Fixed component ID

  5655OPE00

Applicable component levels

• R310 PSY UK27357

     UP07/07/26 P F707

• R410 PSY UK27358

     UP07/07/26 P F707

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.