Opened 16 years ago

Last modified 11 years ago

#282 new defect

IZ1836: qmaster packs two sigkill in one packbuffer for one job

Reported by: roland Owned by:
Priority: normal Milestone:
Component: sge Version: 6.0u6
Severity: Keywords: Sun cleanup


[Imported from gridengine issuezilla]

        Issue #:      1836             Platform:     Sun      Reporter: roland (roland)
       Component:     gridengine          OS:        All
     Subcomponent:    cleanup          Version:      6.0u6       CC:    None defined
        Status:       NEW              Priority:     P3
      Resolution:                     Issue type:    DEFECT
                                   Target milestone: ---
      Assigned to:    ernst (ernst)
      QA Contact:     ernst
       * Summary:     qmaster packs two sigkill in one packbuffer for one job
   Status whiteboard:

     Issue 1836 blocks:
   Votes for issue 1836:

   Opened: Thu Oct 20 02:38:00 -0700 2005 

The bug is described in detail in Issue 1679

In case of tight integration slave tasked killed (coverd by testsuite) the
qmaster sends in one packbuffer two SIG_KILL to the master task. It seems this
is caused because the qmaster mark the job as deleted because one execd reports
one slave died and in the same time another execd sends the load report
intervall with the job as running.

The qmaster should check if the job is already signaled in one packbuffer and
should discard the second kill

Change History (0)

Note: See TracTickets for help on using tickets.