Dnia [Mon, 22 Jun 2009 at 11:14:45AM +0100], Miguel Oliveira
napisał(a):
> Hi all,
>
> We at LIP-Coimbra are managing a hybrid cluster (local+grid usage).
> At the moment one of our local users has been trying to submit a
> large number of jobs (>3000) and the queuing system grinds to a halt
> with all the consequences that has. We have had from extremely slow
> response times on all commands (showq, qstat, qsub, etc) to torque
> and/or maui crashing.
I have seen this with PBS Pro.
Pls post output from
qstat -as
and
tracejob
on the 1st waiting job from those 3000 jobs.
--
Pawel Dziekonski <[log in to unmask]>
Wroclaw Centre for Networking & Supercomputing, HPC Department
Politechnika Wr., pl. Grunwaldzki 9, bud. D2/101, 50-377 Wroclaw, POLAND
phone: +48 71 3202043, fax: +48 71 3225797, http://www.wcss.wroc.pl
|