[torqueusers] limiting resource usage with torque
listsarnau at gmail.com
Fri Jan 20 02:28:12 MST 2012
On Fri, 20 Jan 2012 13:44:57 +1100
Christopher Samuel wrote:
> > After some debugging we found the source. MAUI was reserving 6gb of
> > mem for each job. so, 4 jobs*6gb of mem = 24gb. All the mem was
> > reserved for those 4 jobs and the node is not selected for running
> > more.
> I'm a bit puzzled as to why you think this is a problem - your jobs
> were requesting 6GB vmem (swap) each and your node has 24GB swap and
> you didn't set a queue limit on pvmem to stop jobs requesting 6gb
> pvmem from being requested.
I don't know if I have understood you or not.
My point is that limiting is not the same as reserving.
So, if I want torque to limit resource usage (6gb in this case) I don't
want torque to tell MAUI to reserve 6 GB for that job.
With the reservation, only 4 jobs (4*6=24) can run, but, if jobs behave
correctly and they don't use more than 2 o 3 GB, we can run up to 8
jobs. So, I'm trying to tell torque: "ei! if the job usese more than
6gb, kill it"...
So, my problem comes from the understanding of limiting/reserving
(which are very diferent concepts).
Defines the resources that are required by the job
and establishes a limit to the amount of resource that can be consumed
** Or the problem is in my conf.
More information about the torqueusers