PQ21659: WTRQ CORRUPTION LEADING TO LOOP IN EVENT WRITER PROCESSING THE SAME EVENT OVER AND OVER.

APAR status

• Closed as program error.

Error description

• Some users have experienced a corruption of the TRACKER
  SUBSYSTEM WTRQ in ECSA where the chain pointer of the first
  queue element points to itself.  This causes a loop in the EVENT
  WRITER task attempting to process the broken queue.
  .
  As the EVENTWRITER passes hundreds of copies of the same event
  to the CONTROLLER, it may experience problems handling the
  volume of invalid data, leading to many different problems, and
  possibly forcing the NMM to shut down.
  .
  While the cause of this problem is yet to be determined, code
  has been developed to detect the corrupted WTRQ and capture
  diagnostic information.  The failing TRACKER started task will
  also be terminated and the WTRQ cleared so that a restart of
  the TRACKER should correct the problem.  This code is currently
  in SMP/E UMOD format, and will be repackaged as a formal PTF.
  .
  CIRCUMVENTION:
      If this problem should occur prior to availability of the
  PTFs for this APAR, the user should STOP the looping TRACKER
  STARTED TASK, then restart it using the BUILDSSX(REBUILD)
  keyword in the TRACKER OPCOPTS.  If the TRACKER does not respond
  to a P OPCx command, it may be necessary to CANCEL the address
  space.
      After the TRACKER has been restarted, the BUILDSSX() keyword
  should be immediately removed from the init parms to prevent
  loss of jobtracking information on future restarts.
  

Local fix

• See CIRCUMVENTION, above.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of OPC systems.                    *
  *                                                              *
  *                 FUNCTION=EW Event Writer                     *
  ****************************************************************
  * PROBLEM DESCRIPTION: Some OPC customers have experienced     *
  *                      a loop in the Tracker's subsystem       *
  *                      WTRQ, the Event Writer queue.           *
  *                      This APAR introduces a trap that        *
  *                      is supposed to detect the problem       *
  *                      and force an abend for diagnostic       *
  *                      purposes.                               *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A problem in the Event Writer queue, a loop in the chain
  pointers of the first queue element thus pointing to itself,
  has occurred in a few occasions to some OPC customers.
  The cause of this problem is still unknown and the introduction
  of this trap has two main purposes: limit the damages for the
  customers by intercepting the loop at the beginning and help
  the collection of useful documentation at the right moment.
  

Problem conclusion

• The Event Writer task has been updated to force an abend
  and a dump if an event returned by the OPC queue routine is
  found to point to itself as the next event.
  The abend is an S0C1 abend in the EQQEWEVH module at
  offset +318.
  Please save the dump created by this abend.
  The Event Writer will terminate when this error occurs.
  It is expected that a restart of the Event Writer will be
  successful.
  100Y
  200Y
  EQQEWEVH
  

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:

  UQ24459 UQ24460 PQ22427

Modules/Macros

• EQQEWEVH
  

Fix information

• Fixed component name

  TME 10 OPC V2 R

• Fixed component ID

  5697OPC01

Applicable component levels

• R100 PSY UQ24459

     UP98/12/28 P F812

• R200 PSY UQ24460

     UP98/12/28 P F812