Opened 17 years ago

Last modified 9 years ago

#72 new defect

IZ405: event protocol: no JATASK_DEL event sent for last task

Reported by: joga Owned by:
Priority: low Milestone:
Component: sge Version: current
Severity: Keywords: cleanup
Cc:

Description

[Imported from gridengine issuezilla http://gridengine.sunsource.net/issues/show_bug.cgi?id=405]

        Issue #:      405              Platform:     All       Reporter: joga (joga)
       Component:     gridengine          OS:        All
     Subcomponent:    cleanup          Version:      current      CC:    None defined
        Status:       NEW              Priority:     P4
      Resolution:                     Issue type:    DEFECT
                                   Target milestone: ---
      Assigned to:    andreas (andreas)
      QA Contact:     andreas
          URL:
       * Summary:     event protocol: no JATASK_DEL event sent for last task
   Status whiteboard:
      Attachments:

     Issue 405 blocks:
   Votes for issue 405:


   Opened: Tue Oct 29 02:53:00 -0700 2002 
------------------------


When an array job is processed, no JATASK_DEL
event is sent for the last array task.

Instead the JOB_DEL event contains the last task
id in addition to the job id.

Example for an array job with 10 tasks:
   311   4254    89. EVENT DEL JATASK 5.1
   312   4254    90. EVENT DEL JATASK 5.2
   313   4254    91. EVENT DEL JATASK 5.3
   314   4254    92. EVENT DEL JATASK 5.4
   315   4254    93. EVENT DEL JATASK 5.5
   316   4254    94. EVENT DEL JATASK 5.6
   317   4254    95. EVENT DEL JATASK 5.7
   318   4254    96. EVENT DEL JATASK 5.8
   319   4254    97. EVENT DEL JATASK 5.9
   320   4254    98. EVENT DEL JOB 5.10

Instead a JATASK_DEL event should be sent for the
last task (in the example EVENT DEL JATASK 5.10)
followed by a JOB_DEL event for the job (in the
example EVENT DEL JOB 5).

   ------- Additional comments from joga Tue Oct 29 03:00:35 -0700 2002 -------
The behaviour seems to depend on timing and the order in which the job
reports are received / processed by qmaster.

Probably it only occurs, if all tasks of the array job finish "at the
same time".

   ------- Additional comments from andreas Mon Mar 8 10:36:43 -0700 2004 -------
Moving to 'cleanup'. It is unclear to what extend this
actually can be observed by end users resp. what misbehaviour
results from it.

Change History (0)

Note: See TracTickets for help on using tickets.