[torqueusers] Help! In Mac OS X All Jobs Are Being Sent to One Node

Steven Saunders s_j_nevets at yahoo.com.au
Fri Mar 11 01:38:34 MST 2005


Hi All,

I have a small problem with torque that I hope someone out there can
help me fix:

I have just set up torque on a small cluster of Mac OS X systems. I
have one system which runs pbs_server and pbs_sched, and a couple of
separate systems running pbs_mom. I basically followed the torque
quick start guide and I've gotten to the point where I can use qsub
to submit a job, and it will run on the first execution node, and the
results are successfully delivered back to the system where the job
was submitted. 

My problem is that when I submit several jobs using qsub, they are
all launched immediately on the first execution node (even if I
submit 50 jobs.) The other execution nodes don't receive any of the
jobs, and all of the jobs are launched simultaneously on the first
execution node.

Each execution node has 2 cpus, so what I'd like to happen is that
jobs 1 and 2 go to node 1, jobs 3 and 4 go to node 2 etc. 

Reading the OpenPBS admin guide, I noticed there are ideal_load and
max_load settings for the MOM config file. Should I be using these,
or is the problem somewhere else? At the moment my MOM config files
contain only the lines recommended in the quick-start guide:

        $clienthost     <my server's IP>
        $logevent       255
        $restricted     <my server's IP>

Thanks in advance for any help anyone can provide.


Find local movie times and trailers on Yahoo! Movies.
http://au.movies.yahoo.com


More information about the torqueusers mailing list