PH70409: ABN=0C4-00000004,LOC=CSQPLPLM.CSQPRAH2+OFFSET AND ABN=5C6-00D70120,U=XXXXX ,C=MQ900.940.BUFF-CSQP3GET,M=CSQGFRCV,LOC=CSQPLPLM.

APAR status

• Closed as program error.

Error description

• The S5C6-00D70120 dump occurred in CSQP3GET when the QMGR
  readahead task couldn't obtain a buffer (for the reported case,
  the buffer pool was sized at its minimum size of 100 buffers) to
  read a page from a page set. The task had repeatedly synced with
  the deferred write processor to wait for it to write out buffers
  on the deferred write queue, but the process had not succeeded
  in writing out any buffers.
  
  Following the S5C6-00D70120 dump, recovery in CSQPRAH2 got
  control and called CSQPMTRM to cleanup the pending I/O. After
  returning from CSQPMTRM, CSQPRAH2's recovery routine abended
  S0C4 due to a register not being saved correctly when calling
  CSQPMTRM.
  
  Since the abend occurred on a QMGR SRB and none of the recovery
  routines retried, recovery in CSQVEUS2 decided to terminate the
  QMGR.
  
  This APAR was raised to improve the processing in this scenario.
  

Local fix

• Increase the number of buffers allocated to the buffer pool wit
  insufficient buffers.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of IBM MQ for z/OS Version 9       *
  *                 Release 4 Modification 0.                    *
  ****************************************************************
  * PROBLEM DESCRIPTION: Abnormal queue manager termination      *
  *                      S6C6- following abend 5C6-00D70120 in   *
  *                      CSQP3GET when there are insufficient    *
  *                      buffers for readahead processing.       *
  *                      In some cases abend 0C4 can also occur  *
  *                      in CSQPRAH2.                            *
  ****************************************************************
  During readahead processing, when the readahead task attempted
  to get buffers for a batch of pages needing to be read in, there
  were no stealable buffers available, even after triggering
  deferred write processing to try and make more available,
  leading to the reported abend 5C6-00D70120 in CSQP3GET. Recovery
  for the abend percolated, leading to the task terminating
  unexpectedly and leading to abnormal queue manager termination.
  An error in recovery processing while cleaning up following the
  abend can also lead to abend 0C4 in CSQPRAH2.
  

Problem conclusion

• CSQPRAH2 has been changed to correct the error in recovery
  processing, and to retry following the expected 5C6-00D70120
  abend, allowing the readahead processor to continue processing
  other buffer pools and preventing the abnormal queue manager
  termination.
  

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:

  UO07841

Modules/Macros

• CSQPRAH2
  

Fix information

• Fixed component name

  IBM MQ Z/OS V9

• Fixed component ID

  5655MQ900

Applicable component levels