[torqueusers] Maui probs

skip at pobox.com skip at pobox.com
Thu Dec 24 11:37:29 MST 2009


Sorry to bother this list with what I think is a Maui questions, but
I've so far been unable to subscribe to the mauiusers mailing list, and
my message to help at supercluster.org has gone unanswered.  I hope someone
here can help me out.

I fired up a few thousand batch jobs via qsub last night.  This morning,
after standing reservations should have excluded new jobs from running
(in my mind at least), I marked all nodes offline:

    sudo pbsnodes -o $(pbsnodes -a | egrep '^[a-z]')

That worked fine.

However, now I want to start things up again.  I cleared the offline
flag for a bunch of nodes:

    sudo pbsnodes -c huron tuba ruth ...

and sure enough, pbsnodes -a shows them as free, for example:

    tuba.wacker
         state = free
         np = 1
         ntype = cluster
         status = opsys=solaris7,...

Unfortunately, I don't see any running jobs:

    % qstat | wc -l ; qstat | egrep ' R ' | wc -l
        2601
           0

I would have thought free nodes could start hosting jobs again.  I must
be misunderstanding something.  Any suggestions appreciated.

-- 
Skip Montanaro - skip at pobox.com - http://www.smontanaro.net/


More information about the torqueusers mailing list