[GE users] Making Very Important job to run
seandavi at gmail.com
Tue Jul 13 14:11:43 BST 2010
[ The following text is in the "utf-8" character set. ]
[ Your display is set for the "ISO-8859-10" character set. ]
[ Some characters may be displayed incorrectly. ]
On Tue, Jul 13, 2010 at 9:02 AM, sgerns <rajansrivastava83 at gmail.com<mailto:rajansrivastava83 at gmail.com>> wrote:
I have few queries in perspective of job priorities & making very Important job to run even if cluster is full .
Scenario: I have all the slots of execution hosts in the cluster which are full with jobs. I want to submit and run very important job on the cluster when all the execution hosts are full.
Is there any way I can do that so that SGE would allow very imporatant job to run.
& other jobs can be put on hold & give space to this important job to run
I don wanna to kill any job & I want this Important job to run immediately.
I have a queue for this job which is having the hisghest priority, but that will make me up in the queue
but wont allow me to run untill all the resources are occupied.
Look into subordinate queues. We have two queues, one subordinate to the other. All users run in the subordinate queue and when we need to run a job immediately, we submit it to the other, non-subordinate queue. Approximately the same thing can be accomplished by having one queue run with nice 10 or something like that and a second queue run with normal nice level.
More information about the gridengine-users