OA58616: PREDICTIVE FAILURE ANALYSIS ABEND0C4 RC10 AIRHMMAP PIC10 PFA HIGH CPU

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• When processing exclude/included jobs, AIRHMMAP loops through
  RTD events.  It uses a 1-byte index (eventIx) to look at these
  events but the count of events can be up to 4 bytes.  This
  causes AIRHMMAP to index past the end of the RTD events, and
  reference storage that is not getmained resulting in the
  Abend0C4.  AIRHMMAP was at base HBB77B0 and the abend0C4 occured
  at offset +x'827A'.
  
  If AIRHMMAP does not Abend, this loop could potentially be
  infinite causing HIGH CPU in the PFA address space.
  
  In a dump of PFA, experiencing high cpu you may see EXT TMR
  entries below in AIRHMMAP, ISGLOBTS, ISGLRELS repetatively in
  the systrace.  offsets in AIRHMMAP can vary, +x'83AC',
  +x'802C', x'82C6'.
  
  example:
  EXT   TIMR 00000000_2DE7E5F4     <--- AIRHMMAP
             07241000 80000000
  EXT   TIMR 00000000_011A23BC     <--- ISGLOBTS
             07240000 80000000
  EXT   TIMR 00000000_011A2612     <--- ISGLOBTS
             07241000 80000000
  EXT   TIMR 00000000_133305F4     <--- ISGLRELS
             07241000 80000000
  

Local fix

• Note: Only one of the following suggestions is necessary to
        circumvent the high cpu, but either PFA checks will be
        unavailable or RTD will be disabled.
  
  1. Disable PFA checks that use data from Runtime Diagnostics.
  
      change parm to stop collection of data for the ENR and
      MAR checks.
      f hzsproc,update,check(IBMPFA,PFA_ENQUEUE_REQUEST_RATE),
      parm('collectinactive(0)')
  
      f hzsproc,update,check(IBMPFA,PFA_MESSAGE_ARRIVAL_RATE),
      parm('collectinactive(0)')
  
      Next, deactivate the checks.  For example:
      f hzsproc,deactivate,check(IBMPFA,PFA_ENQUEUE_REQUEST_RATE)
  
  2. Stop the Runtime Diagnostics (HZR address space).
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * All users of HBB77B0 and HBB77C0,                            *
  * exploiting the Predictive Failure                            *
  * Analysis (PFA).                                              *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * Predicative Failure Analysis (PFA) received ABEND0C4 PIC10   *
  * in AIRHMMAP.                                                 *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  PFA_ENQUEUE_REQUEST_RATE or  PFA_MESSAGE_ARRIVAL_RATE  checks
  while processing for excluded jobs, calls
  Find_Excluded_Job_Events routine in AIRHMMAP and loops
  through events reported by RunTime Diagnostics.  The index for
  looping was erroneous resulting in ABEND0C4.
  

Problem conclusion

• Code has been changed to prevent the abend condition.
  

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:

  UJ01763 UJ01764 UJ01774

Modules/Macros

• AIRHMMAP
  

Fix information

• Fixed component name

  Z/OS PFA

• Fixed component ID

  5752SCPFA

Applicable component levels

• R7B0 PSY UJ01764

     UP20/01/22 P F001

• R7C0 PSY UJ01774

     UP20/01/22 P F001

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.