[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] Whole System Scheduling
- Date: Fri, 23 Oct 2009 13:28:59 -0400
- From: "Jonathan D. Proulx" <jon@xxxxxxxxxxxxx>
- Subject: Re: [Condor-users] Whole System Scheduling
On Fri, Oct 23, 2009 at 09:18:02AM -0700, Ian Chesal wrote:
:You mean like dynamic provisioning? In 7.2.x this is available. I myself haven't tried it yet. But the docs say it's there. See:
That may just make my day! I am currently running 7.2, guess I
haven't been reading the release notes as closely as I might have.
A quick look make me think that there's no way to prevent jobs with
high resourse needs from being starved out by lower need jobs. For
example if there's a large number of low priority single CPU jobs such
that the cluster is fully partitioned a single high priority two cpu
job will never run?
I'll try it ou tand report back on how it works for me.