IZ2913: delayed job deletion when spooling on NFS

[crei]

[Imported from gridengine issuezilla ​http://gridengine.sunsource.net/issues/show_bug.cgi?id=2913]

        Issue #:      2913             Platform:     Sun      Reporter: crei (crei)
       Component:     gridengine          OS:        All
     Subcomponent:    execution        Version:      6.2         CC:    None defined
        Status:       NEW              Priority:     P4
      Resolution:                     Issue type:    DEFECT
                                   Target milestone: ---
      Assigned to:    pollinger (pollinger)
      QA Contact:     pollinger
          URL:
       * Summary:     delayed job deletion when spooling on NFS
   Status whiteboard:
      Attachments:

     Issue 2913 blocks:
   Votes for issue 2913:


   Opened: Fri Feb 13 05:30:00 -0700 2009 
------------------------


This problem occured when running testsuite test 2743 on CSP classic spooling
with the option no_local_spool.

The test is submitting a 400 second sleeper job to a execd.
The job paramters are "-l h_rt=30 -o /dev/null -j y -b y /bin/sleep 400"

after that the execd is shutdown
after 30 seconds the execd is restarted again

Job should be deleted by execd at startup, but when qmaster and execd are
spooling to a NFS directory it takes one minute until the job is vanished.