PH57260: MQ Z/OS: CSQV086E REASON=00F30908 OCCURRED UNEXPECTEDLY DUE TO A LOGIC ERROR

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• REASON=00F30908 was added by APAR PH27803 to indicate that End
  of Memory and Termination processing detected an unexpected
  loop in the VLCAAACE control block chain.
  
  When an address space ends, IBM MQ has processing to ensure
  that all connections have completed before cleaning up the
  state relating to the address space. As part of this
  processing, there is a loop which checks the state of the
  remaining connections (ACE control blocks) and waits for
  1 second before checking again.
  
  In the reported case, there was not a control block loop.
  Rather, there was a job that had been hung for days due to the
  problem fixed by APAR PH50958.  Because the hung thread did not
  end and the queue manager was not restarted, the 1-second loop
  continued.
  
  Within this main loop is an inner loop that runs a chain of ACE
  control blocks to check the state of each remaining block.
  There is a counter in the inner loop that is intended to
  detect a loop in the chain and abend the queue manager if we
  process more than 5,000,000 (5 million) elements in the chain.
  
  For the intended loop-detection behavior, the loop counter
  should be set to zero before we start the inner loop each time.
  However, the initialization of the counter is wrongly done
  before the start of the outer loop. Since the hung thread means
  that we are repeatedly going round the outer loop once a
  second, the loop counter gradually increases over time and
  reaches the defined limit after 57 days. MQ incorrectly
  believes that there is a loop in the ACE chain and abends the
  queue manager with reason 00F30908.
  
  The loop-checking logic is not intended to detect a hung
  thread, so the abend is not appropriate in those circumstances.
  

Local fix

• Until PH50958 can be applied, if there is a hung job, recycle
  the queue manager before 57 days.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of IBM MQ for z/OS Version 9       *
  *                 Release 2 Modification 0 and Release 3       *
  *                 Modification 0.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: If a thread is hung in MQ, then under   *
  *                      certain circumstances the QMGR could    *
  *                      terminate with reason 00F30908          *
  *                      approximately 57 days later. This will  *
  *                      be accompanied by a dump being          *
  *                      scheduled.                              *
  ****************************************************************
  Incorrect checking can result in a hung thread being incorrectly
  interpreted as a loop in a control block chain. The QMGR
  responds to this scenario by terminating the QMGR with reason
  00F30908.
  

Problem conclusion

• The control block loop detection code has been corrected to
  prevent the QMGR from terminating.
  

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:

  UI94299 UI94300

Modules/Macros

• CSQ3SSI1 CSQ3SSI2
  

Fix information

• Fixed component name

  IBM MQ Z/OS V9

• Fixed component ID

  5655MQ900

Applicable component levels

• R200 PSY UI94300

     UP23/11/15 P F311 ¢

• R300 PSY UI94299

     UP23/11/15 P F311 ¢

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.