OA47327: PURGEDQ SRB ABEND47B IXGWRITE ABEND017 ABEND1C5 IXGA1MTX

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• For STOP command processing, a PURGEDQ was issued against SRBs
  running in the stopped address space. An address space SRB
  issuing IXGWRITEs had been suspended  by IXGF2WRT for
  completion of a staging data set I/O request sent to Media
  Manager. The system abended the SRB with an ABEND47B. IXGF2WRT
  recovery retried the abend47b and returned control to the
  application with Return Code x'C' indicating a component error.
  The application SRB continued to run issuing subsequent
  IXGWRITEs. Subsequent IXGF2WRT attempts to SUSPEND for I/O
  completion were not successful and were returned RC=8 due to the
  earlier abend47b on the SRB.  The logic in logger for DasdOnly
  writes records the error but then continues as if the async I/O
  to the staging data set had completed which is untrue in some
  cases.
  Processing in IXGF2wrt after the suspend rc8  will release the
  storage related to the async I/O, i.e. RQE, Arte, RWIO buffer
  list, etc.  which should not happen until it is clear that the
  I/O is complete and IXGA1MTX  (Media Manager Termination Exit)
  processing is finished for the particular I/O request. This can
  cause subsequent failures when IXGA1MTX is called by media
  manager for I/O completion since the coordination between System
  Logger and Media Manager for asynchronous I/O becomes disrupted.
  

Local fix

• -Force and Restart IXGLOGR if problem is encountered
  -See scatch pad for application workaround
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: Users of HBB7780 and above supported         *
  *                 releases using System Logger Service         *
  *                 IXGWRITE in SRB mode.                        *
  *                 Also see APAR OA47836.                       *
  ****************************************************************
  * PROBLEM DESCRIPTION: ABEND017 rsn20C on attempt to resume    *
  *                      SRB thread that is issuing IXGWRITE     *
  *                      requests after this SRB has previously  *
  *                      been targeted for cleanup via PURGEDQ   *
  *                      as part of the logstream exploiter      *
  *                      address space stop processing           *
  *                                                              *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  An application connected to a DasdOnly logstream had an SRB
  issuing IXGWRITE mode=sync requests.  One write request had been
  suspended in the front end processing of logger module IXGF2WRT.
  The suspension occurred beause the module was waiting for
  completion of the asynchronous staging data set I/O request sent
  to DFSMS Media Manager (as expected).
  
  While the SRB was suspended, a PURGEDQ was issued and targeted
  the SRBs in this application's address space.  This caused
  IxgF2wrt front end processing to receive the ABEND47B condition.
  IXGF2WRT recovered, retried via RTM, and returned control to the
  application with Return Code x'C' Reason Code x'0000' indicating
  a component error.
  
  The application SRB continued to run issuing subsequent
  IXGWRITEs.  These subsequent IXGF2WRT attempts to SUSPEND for
  I/O completion were not successful as Logger received Return
  Code x'8' due to the earlier ABEND47B on this SRB.
  
  The logic in logger for DasdOnly writes records the error, but
  then continues as if the async I/O to the staging data set had
  completed, which is untrue in some cases.  Processing in
  IXGF2wrt after the suspend rc8 will release the storage related
  to the async I/O, i.e. RWIO buffer list, Arte, RQE, etc.  The
  resource cleanup should not happen until it is clear that the
  I/O is complete.
  
  The premature releasing of the resources can cause subsequent
  failures in Logger and/or DFSMS Media manager since the
  coordination between System Logger and Media Manager for
  asynchronous I/O becomes disrupted.
  For example, an ABEND017 rsn20c can occur because of an attempt
  to resume an SRB using a bad SPTOKEN at IXGA1MTX UA70736+1004
  ?RESUME SPTOKEN(Rqe_SpToken).  This occurs because the RQE had
  been erroneously released by the IxgF2wrt recovery/retry
  processing.
  
  Logger should be corrected to properly handle the cleanup of the
  resources associated with the asynchronous concurrent threads.
  Also, when the logger write front end processing receives an
  ABEND47B, it should cleanup appropriately and percoloate to the
  IXGWRITE invoker.  The percolation will allow the invoker to
  encounter the ABEND47B and handle it as needed.
  

Problem conclusion

• This fix for APAR OA47327 provides a corrective PTF only on
  release level HBB7790.  Note the parts included in this PTF are
  not shipped in HBB779J.  See Logger APAR OA47836 for corrective
  PTF(s) on other release levels.
  
  Logger has been updated to percolate ABEND47B conditions that
  are encountered while in module IXGF2WRT front end processing
  (i.e.  running in caller's SRB environment).
  
  Corrections have also been made in the handling of the
  asynchronous related control blocks so that they are released
  when properly owned by the concurrent threads.
  
  
  This fix may be applied to one system at a time in the sysplex.
  
  Systems with and without the fix activated can coexist in the
  same sysplex, although the problem may persist on those systems
  running without the fix.
  
  
   Additional Searchable keywords:
     SYSPLEXDS
  

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:

  UA77313

Modules/Macros

• IXGA1MTX IXGC4DIS IXGF1CMP IXGF2WRT IXGL2RTE
  IXGR2RMT IXGV1FMT IXGYCFMT
  

Fix information

• Fixed component name

  SYSTEM LOGGER

• Fixed component ID

  5752SCLOG

Applicable component levels

• R790 PSY UA77313

     UP15/05/27 P F505

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.