IZ2783: Enhance shutdown behaviour of new IJS when client is stopped with CTRL + C or killed with SIGKILL

[crei]

[Imported from gridengine issuezilla ​http://gridengine.sunsource.net/issues/show_bug.cgi?id=2783]

        Issue #:      2783             Platform:     All          Reporter: crei (crei)
       Component:     gridengine          OS:        Windows XP
     Subcomponent:    execution        Version:      6.2             CC:    None defined
        Status:       NEW              Priority:     P3
      Resolution:                     Issue type:    DEFECT
                                   Target milestone: Maintrunk
      Assigned to:    pollinger (pollinger)
      QA Contact:     pollinger
          URL:
       * Summary:     Enhance shutdown behaviour of new IJS when client is stopped with CTRL + C or killed with SIGKILL
   Status whiteboard:
      Attachments:

     Issue 2783 blocks:
   Votes for issue 2783:


   Opened: Mon Nov 10 04:11:00 -0700 2008 
------------------------


On our windows host the shepherd of a qrsh stays in deleting state for 1 Minute
after client was stopped with CTRL + C. The trace file of the job reports
following messages:

Trace file contains (for CTRL+C and for killed qrsh):
      11/07/2008 12:07:33 [197691:2185]: sending UNREGISTER_CTRL_MSG with
exit_status = "137"
      11/07/2008 12:07:33 [197691:2185]: sending to host: hostfoo
      11/07/2008 12:07:33 [197691:2185]: waiting for UNREGISTER_RESPONSE_CTRL_MSG
      after some time (1 minute)
      11/07/2008 12:08:33 [197691:2185]: Server already exited
      11/07/2008 12:08:33 [197691:2185]: main: cl_com_ignore_timeouts
      11/07/2008 12:08:33 [197691:2185]: main: leaving closinge_parent_loop()

At least the shepherd disappears from the queue and terminates, but the shutdown
behavior should be enhanced to be a little bit faster.

   ------- Additional comments from crei Mon Nov 10 04:11:54 -0700 2008 -------
Setting target milestone to maintrunk