IZ92: enhance shutdown functionality

[crei]

[Imported from gridengine issuezilla ​http://gridengine.sunsource.net/issues/show_bug.cgi?id=92]

        Issue #:      92              Platform:     All           Reporter: crei (crei)
       Component:     hedeby             OS:        All
     Subcomponent:    bootstrap       Version:      1.0              CC:    None defined
        Status:       NEW             Priority:     P3
      Resolution:                    Issue type:    ENHANCEMENT
                                  Target milestone: 1.0u5next
      Assigned to:    adoerr (adoerr)
      QA Contact:     adoerr
          URL:
       * Summary:     enhance shutdown functionality
   Status whiteboard:
      Attachments:


     Issue 92 blocks:
   Votes for issue 92:     Vote for this issue


   Opened: Tue Aug 21 03:14:00 -0700 2007 
------------------------


   Currently the hedeby shutdown only triggers the shutdown of hedeby components.

   It would be better to wait until the components are really down and after
   that return from the sdmadm shutdown call. The command should at least
   wait a specified timeout for the components to shut down, if they don't
   go down - kill the components.

   If an administrator triggers shutdown and does immediate a restart of a
   component the current behavior might raise race conditions and problems.

   If all component should be shutdown the configuration service is the last one
   which is shutdown. Also here the shutdown command should wait that all
   components had time to go down before shutting down CS. Currently
   the shutdown command waits 5 seconds and assumes that all components had enough
   time to go down.
               ------- Additional comments from easymf Tue Nov 27 10:31:44 -0700 2007 -------
   will look into this
               ------- Additional comments from easymf Tue Nov 27 10:33:47 -0700 2007 -------
   - waiting for the shutdown of the JVMS is possible (i have prepared working
   concept in my workspace).
   - killing of the JVMS is not possible because the command works with remote JVMS
   (killing of the JVM on localhost is already done)
               ------- Additional comments from crei Wed Dec 5 06:31:45 -0700 2007 -------
   not sure if this should be fixed for code freeze
               ------- Additional comments from easymf Tue Aug 5 05:22:23 -0700 2008 -------
   changing the task to enhancement of version 1.
               ------- Additional comments from easymf Tue Aug 5 05:23:07 -0700 2008 -------
   priority is definitely lower that p1
               ------- Additional comments from rhierlmeier Wed Nov 25 07:21:10 -0700 2009 -------
   Milestone changed