PM71339: JOBS REMAIN IN SUBMITTED STATE UNTIL ENDPOINTS ARE RECYCLED.

8.0.0.3: WebSphere Extended Deployment Compute Grid V8.0 Fix Pack 3
8.0.0.4: WebSphere Extended Deployment Compute Grid V8.0 Fix Pack 4
8.0.0.5: WebSphere Extended Deployment Compute Grid V8.0 Fix Pack 5

APAR status

• Closed as program error.

Error description

• Sometimes jobs in customer environment remain in the submitted
  state, despite indications from joblogs or other log files that
  they should have reached some other state (ended, become
  restartable, etc.).  They have to interrupt their service by
  cycling endpoints, the job scheduler, or sometimes even the cell
  to clear the issue.
  

Local fix

• For the time being customer can recycle the endpoints to clear
  the issue.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED:  All users of WebSphere Extended Deployment  *
  *                  Compute Grid Version 8                      *
  ****************************************************************
  * PROBLEM DESCRIPTION: Problem initializing job log            *
  *                      processing has side effect leading to   *
  *                      jobs stuck in submitted state.          *
  *                      A second problem is that jobs that      *
  *                      failed before their first step ran      *
  *                      were, on restart, going                 *
  *                      right to "ended" state without          *
  *                      executing any of the jobs' steps.       *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A problem initializing job log handling for a given job was
  handled incorrectly by the runtime
  on the endpoint server.  This led to a problem downstream
  where the job status wasn't getting
  communicated and managed correctly.  So rather than the job
  appearing to have failed and
  been put into "restartable" state, the job appears to be stuck
  in "submitted" state (still
  in "submitted" state since job log initialization happens
  early on in the job lifecycle).
  A second problem found at the same time occurs when a job
  fails and is put into "restartable"
  state before the first step ever gets executed.   The problem
  appears when the job is
  restarted.   The restarted job appears to execute successfully
  and finish in the 'ended'
  state, according to the external job status as seen for
  example in the job management
  console.   However, none of the steps comprising the job will
  actually have executed, due
  to a flaw in the logic used by the batch container in restart
  execution.
  

Problem conclusion

• Job log initialization error handling has been tightened up so
  that a failure puts the
  job into restarted state.   The bug in restart execution has
  been fixed so that upon
  restart, a job that failed before the first step was attempted
  to be executed will
  resume execution at the first step.
  The fix for this APAR is currently targeted for inclusion in
  fixpack 8.0.0.3. Please refer to the Recommended Updates page
  for delivery information:
  http://www.ibm.com/support/docview.wss?uid=swg27022998
  

Temporary fix

• An interim fix is available upon request.
  

Comments

APAR Information

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

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

  PM79241

Fix information

• Fixed component name

  WXD COMPUTE GRI

• Fixed component ID

  5725C9301

Applicable component levels

• R800 PSY

     UP