[GE users] Strange qstat -g c -ext Vs qstat -qs E

Amit H Kumar AHKumar at odu.edu
Fri Aug 4 19:19:29 BST 2006



Sean Dilda <sean at duke.edu> wrote on 08/04/2006 02:07:59 PM:

> Amit H Kumar wrote:
>
> >
> > Still I want to answer a basic question:
> >       How do I tell if the error was due to a job only. if it is
possible.
> >
>
> As far as I know, the only way a queue will go into error state is due
> to a job failure.   When a job fails, SGE looks at why.  For certain
> failure reasons, SGE will assume that the job is at fault, and put the
> job in error state.  For certain other reasons, it will assume there's
> something wrong with the queue/host and put the queue in error state.
> Unfortunately, its not always correct in its logic.  Then again, its
> impossible for SGE to always be right  (if it can't access the working
> directory for a job, it might be because the node is missing an NFS
> mount point, or it might be because the job requested a working
> directory that doesn't exist).

Thank you Sean, "a plethora of infromation to chew on" :)
I think I get the idea.
Thank you,
-AK

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list