[torqueusers] pbsnodes offlline->online
may at chg.ru
Wed Aug 22 12:23:03 MDT 2007
Garrick Staples пишет:
> On Wed, Aug 22, 2007 at 10:55:30AM +0400, Anton Menshutin alleged:
>> I set offline for all nodes in order to do some software update, after
>> that i set state back to online, but jobs still hang in queue and does
>> not want to start. New submitted jobs are being run automaticaly, but
>> the old jobs could only be started by manually by qrun.
>> How can it be fixed? I'm running maui as a scheduler.
> maui likely has the jobs held. What does 'checkjob $jobid' say? Or use maui's
> releasehold command, 'releasehold -a ALL'.
Thanks, releasehold really helped. But i don't understand, wheather
this is a bug in maui or it's feature? It is not a first time i see
that if at some time there are no resources available (cause some host
is down - not busy) later batch system refuses to run this jobs.
More information about the torqueusers