[GE users] Using qdel leaves queues in error status

Andreas.Haas at Sun.COM Andreas.Haas at Sun.COM
Tue May 27 16:08:18 BST 2008

Hi Filipe,

On Tue, 27 May 2008, Filipe Brandenburger wrote:

> Hi Andreas,
> Thank for your answer. I guess the best path will be the upgrade to
> 6.1u4, I will use this bug as (one more) reason to give the upgrade some
> priority.

very reasonable. 6.1 resource quotas were reason enough.

> Andreas.Haas at Sun.COM wrote:
>> but note #752 mentions specifically array jobs: Do you have array-jobs?
>> If not, then you may not expect that #752 will bring you an improvement.
> Ok, good to know. We have some array jobs but most of our jobs are
> launched individually, so I don't think this would fix our issue.

That confirms my impression from the logging file you pasted into your initial mail.

>>> I would like to understand what caused this difference in behaviour,
>>> since I don't really like the idea of having processes (specially lots
>>> of them) being killed with SIGKILL.
>> Are you using -notify submit option?
> In some jobs yes, in some jobs no. It really depends on the developers.

You could enforce this  by adding -notify into your

    $SGE_ROOT/$SGE_CELL/common/sge_request file.

> Why? Would that make a difference?

It just reminded me to the notify. When you look into qsub(1) you will 
find that SIGUSR2 is usually delivered as kill notification. If you
use NOTIFY_KILL in sge_conf(5) this could be an explaination of different
signals being delivered.



Sitz der Gesellschaft: Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten
Amtsgericht Muenchen: HRB 161028
Geschaeftsfuehrer: Thomas Schroeder, Wolfgang Engels, Dr. Roland Boemer
Vorsitzender des Aufsichtsrates: Martin Haering

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