[torquedev] max_user_queuable broken in 2.4.8 ?

David Beer dbeer at adaptivecomputing.com
Wed Dec 29 08:31:37 MST 2010


Chris,

Perhaps Glen will find something different, but I was unable to reproduce the problem:

$ qmgr -c 's q batch max_user_queuable=20'
$ ../../sub_jobs.sh 21
0.napali
1.napali
2.napali
3.napali
4.napali
5.napali
6.napali
7.napali
8.napali
9.napali
10.napali
11.napali
12.napali
13.napali
14.napali
15.napali
16.napali
17.napali
18.napali
19.napali
qsub: Maximum number of jobs already in queue for user MSG=total number of current user's jobs exceeds the queue limit: user dbeer at napali, queue batch

If Glen is also unable to reproduce things, I would check that all of the queues have the limit applied, and make sure that it isn't somehow circumvented by a routing queue.

Cheers,

David

----- Original Message -----
> Hi folks,
> 
> We're currently running 2.4.8 (no chance to upgrade for
> the forseable future) and we've just had a couple of users
> submit around 40,000 jobs in total.
> 
> Moab 5.4.2 has completely choked on this and has become
> very slow to respond. We can't upgrade to 6.0 yet so I'm
> trying to limit the number of jobs users can submit.
> 
> I can see the Torque docs refer to a queue attribute of
> max_user_queuable and I've set that to 4000 on both our
> queues but I can still submit jobs as one of the offending
> users who has 18,000 jobs in the queue currently.
> 
> # qmgr -c 'p s' | grep max_user_queuable
> set queue batch max_user_queuable = 4000
> set queue smp max_user_queuable = 4000
> 
> [naughty at bruce-m ~]$ echo sleep 60 | qsub
> Warning: you did not specify a shell in the first line of your PBS
> script
> We have assumed you wish to use bash, however please update your
> script with a
> valid shell
> 
> 473550.bruce-m.vlsci.unimelb.edu.au
> [naughty at bruce-m ~]$ qstat 473550
> Job id Name User Time Use S Queue
> ------------------------- ---------------- --------------- -------- -
> -----
> 473550.bruce-m STDIN naughty 0 Q batch
> 
> That message about bash comes from our submit filter, and
> I've changed the username to protect the guilty.
> 
> Any ideas ?
> 
> Chris
> --
> Christopher Samuel Senior Systems Administrator
> VLSCI - Victorian Life Sciences Computational Initiative
> Email: samuel at unimelb.edu.au Phone: +61 (0)3 903 55545
> http://www.vlsci.unimelb.edu.au/
> _______________________________________________
> torquedev mailing list
> torquedev at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torquedev

-- 
David Beer 
Direct Line: 801-717-3386 | Fax: 801-717-3738
     Adaptive Computing
     1656 S. East Bay Blvd. Suite #300
     Provo, UT 84606



More information about the torquedev mailing list