OA57673: ISGXSRBW TIMED WAIT HANG.

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• There is small timing window in which the ISGXSRBW (Timed WAIT)
  service's FRR, running as a result of an abend against a unit of
  work using the Timed WAIT service, may freemain a timed
  wait DIEAREA of a second (different) work unit.  This leads to
  the timed wait not completing under that second work unit and,
  that work unit becoming permanently hung.
  
  In the case of the reporting customer, an asynchronous abend, an
  ABEND04E, was the catalyst leading to the ISGXSRBW FRR running
  within the window.
  
  ADDITIONAL SYMPTOMS:
  - Module ISGXSRBW is also known as IEAVXTSW.
  - This problem has been seen to cause a hang in IEAVTABD during
    RTM user dump processing.
  .
  DIAGNOSTIC STEPS:
  
  1) From an SVC dump of the hung function, identify the TCB or
     SSRB address for the hung unit of work.  Verify that this
     unit of work is waiting to be woken up by service ISGXSRBW.
       - For a  TCB,  check RBOPSW of the newest [top] RB.
       - For an SSRB, check SSRBCPSW.
  
  2) Using the work unit address, get the associated WEB address:
        For a  TCB address  x: IP L x+138?+E8 LEN(4)
        For an SSRB address y: IP L y+20     LEN(4)
  
  3) Format the WEB address z:  IP CBF z STR(WEB)
     and do: FIND TOD.
  
  4) Do:  IP LTOD  w    where w is the doubleword value in the
                        WEB_TOD field, with all blanks removed.
  
     This will provide the time that this work unit last entered a
     wait.  If the difference between this time and the time of
     the dump is significantly more than the time interval set
     using the ISGXSRBW service (typically a short, period of
     time), then you are most likely experiencing this problem.
  
  
  Following verification technique is an alternative to steps 1-4.
  
  - When formatted in IPCS, the hung work (TCB or SRB) has a
    FRRs PARM area second word pointing to a
    'IEAVXTSW DIEAREA'.
  
  - VSM OWNCOMM data would show that area is not owned by
    the hung work ASID. Additionally, the SRB within the
    DIEAREA does not have a purge ASID (PASID) or ASCB
    that matches the hung work.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  *  Users running z/OS HBB77A0 and above                        *
  *                 who use the IEAVXTSW/ISGXSRBW service        *
  *                 (documented in commentary of ECVT field      *
  *                 ECVTXTSW).                                   *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  *  A IEAVXTSW/ISGXSRBW Timed Wait call                         *
  *                      can hang following an asynchronous      *
  *                      abend.                                  *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  Workarea used by IEAVXTSW/ISGXSRBW can incorrectly be
  referenced if an asynchronous abend occurs.
  

Problem conclusion

• Fix workarea references for the case of an asynchronous
  abend occurring.
  

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:

  UA99889 UA99890

Modules/Macros

• IEAVXTSW
  

Fix information

• Fixed component name

  SUPERVISOR CONT

• Fixed component ID

  5752SC1C5

Applicable component levels

• R7A0 PSY UA99890

     UP19/07/24 P F907

• R7B0 PSY UA99889

     UP19/07/24 P F907

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.