[GE users] Preemption

Charu Chaubal Charu.Chaubal at Sun.COM
Fri Oct 29 19:50:11 BST 2004


Hi Brian,

The priority flag mentioned below is the dispatch priority, not the runtime nice
value.

Nevertheless, for pre-emption, you must use subordinate queues.  Reuti's email
link below shows how to override the default pre-emption behavior, which is
"suspension", to instead perform a "kill and resubmit".

Regards,
	Charu


Brian R. Smith wrote:
> Reuti,
> 
> I'm sorry to bother you again, but I'd like to dictate the terms of
> priority for each queue and I thought that priority was the nice level
> of the process (but I guess thats only for the queue_conf).  I just want
> all.q to have lower priority (not nice-level) in all cases, regardless
> of job size.
> 
> Could I have a back-end script that checks to see the requested queue
> and sets the job priority accordingly?
> 
> As for the suspend-method, I will look into that.
> 
> Your help is greatly appreciated.
> 
> Brian
> 
> On Fri, 2004-10-29 at 20:22 +0200, Reuti wrote:
> 
>>Hi,
>>
>><snip>
>>
>>>1) Regardless of the queue priority and the queue used, multi-processor
>>>jobs ALWAYS have higher priority and are run first.  I have a 'group1.q'
>>>and an 'all.q' that is subordinate to 'group1.q'.  If I submit a
>>>parallel job to 'all.q', it MUST have lower priority than any job that
>>>gets submitted to group1.q.  How can this be accomplished?
>>
>>changing the priority to -1023 with qsub is not an option?
>>
>>
>>>2) My two research groups would like to divide the cluster resources in
>>>half, e.g. one group gets 6 nodes and the other group gets the other 6
>>>nodes.  However, they also want the option to run on ALL of the nodes.
>>>The caveat is that if one group's job prevents the other group from
>>>running jobs, the scheduler should remove enough offending jobs from the
>>>queue to allow the rightful group to run their job(s).
>>>
>>>Currently, I have three queues, group1.q, group2.q and all.q.  all.q is
>>>subordinate to both group1.q and group2.q via the queue_conf parameter
>>>subordinate_list.  I would like my group1.q and group2.q queues to
>>>delete any job running in all.q that is preventing a scheduled job to
>>>the two high priority queues from running.  How can this be
>>>accomplished?
>>
>>if I understand the problem in the correct way, you have to use a 
>>suspend-method for all.q which will kill or requeue the job. There was a 
>>discussion about:
>>
>>http://gridengine.sunsource.net/servlets/ReadMsg?msgId=20199&listName=users
>>
>>The only problem may be any defined resources. They are not freed, before the 
>>all.q job is killed, and so preventing from starting the job in group[12].q at 
>>all.
>>
>>Cheers - Reuti
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
>>For additional commands, e-mail: users-help at gridengine.sunsource.net
> 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net
> 

-- 
####################################################################
# Charu V. Chaubal              # Phone: (650) 786-7672 (x87672)   #
# Grid Computing Technologist   # Fax:   (650) 786-4591            #
# Sun Microsystems, Inc.        # Email: charu.chaubal at sun.com     #
####################################################################


---------------------------------------------------------------------
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