PM74848: TASKS CAN HANG IN DEFERRED FCIOWAITS AND FCXCSUSP BECAUSE A BROWSE AGAINST A CMT DATA TABLE HAS NOT ENDED.

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Under certain circumstances, it is possible for a
  STARTBR against a CMT Data Table to result in a VSAM Point to
  the source dataset, and then subsequent READNEXT or READPREV
  are handled through the Data Table without going through
  VSAM.  When that happens, the VSAM Point causes position to
  be held on one of the buffers until ENDBR from the application.
  That can lead to more Exclusive Control locking, and the
  circumstances leading to VSAM APAR OA36624 are met so that
  that problem may happen also.
  

Local fix

• Make the source dataset NSR
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All.                                         *
  ****************************************************************
  * PROBLEM DESCRIPTION: A STARTBR against a CICS Maintained     *
  *                      Data Table can lead to tasks hanging    *
  *                      with deferred FCIOWAITs and FCXCSUSPs.  *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  During a start browse against a CMT Data Table, if DT services
  receives a DTP_NOT_VALID response it will instead issue a start
  browse against the base data set (VSAM POINT). The following
  READNEXT or READPREV requests may not get this DTP_NOT_VALID
  response and so will have no need to attempt the request against
  the base data set. When an update request, such as a delete is
  issued, it will issue a VSAM GET UPD and it may enter an
  FCXCSUSP suspend due to the previous VSAM POINT that the STARTBR
  issued.
  
  In the majority of cases, the update request would have led to
  an AFCG abend because the VSAM GET UPD would have returned with
  0814. However in the reported problem the update request also
  met the circumstances leading to VSAM APAR OA36624. This meant
  that the update request did not abend but suspended waiting for
  the control interval and other tasks built up behind it.
  
  The same situation could occur if a reset browse (RESETBR)
  request received a DTP_NOT_VALID response as well.
  
  Additional Keywords: abendAFCG AFCG
  

Problem conclusion

• DFHFCDTS has been altered so that if a POINT is requested
  against the source as part of a STARTBR or RESETBR request then
  a corresponding ENDREQ will also be issued against the source.
  This will ensure that when browsing CICS Maintained
  Data Tables, VSAM resources are not held unnecessarily.
  This fix also prevents self-deadlock AFCG abends from occurring
  in certain circumstances.
  

Temporary fix

• FIX AVAILABLE BY PTF ONLY
  

Comments

APAR Information

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

  PM72706

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

  UK83194 UK83195

Modules/Macros

•    DFHFCDTS
  

Fix information

• Fixed component name

  CICS TS Z/OS V4

• Fixed component ID

  5655S9700

Applicable component levels

• R600 PSY UK83194

     UP12/11/09 P F211

• R700 PSY UK83195

     UP12/11/09 P F211

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.