The PBS config files only appear on one machine (the head node?); make
sure to check for them on the relevant node. However, the output from
qmgr and qstat means that PBS is configured somehow -- I'm guessing
that you're looking for the config files in the wrong place.
The big question is this:
> > But, if i ask a PBS job in interactive mode, like this :
> >
> > [umr65_at_editr SCRATCH]$ qsub -lnodes=2 -I
> > qsub: waiting for job 43.editr.cluster.ird.nc to start
> >
> > After a long time, PBS still waiting ... If i check with "qstat -f",
> > "Resource_List.ncpus "is always equal to 1. However i asked 2 nodes !
> > What
> > is wrong ? Do you want other log, output ?
Do you ever get a job shell? From this text, it's not clear if your
prior tests were with this same qsub line or a different command line
(because you said that your prior tests *ran*, but unexpectedly only
with one node).
You never answered my questions about what you meant with your problems
with lamd's possibly remaining on a node after the job completed. So
I'm assuming that I either misunderstood the question, or it's somehow
no longer a problem.
So I hand this thread off to the OSCAR list and someone who can answer
PBS questions...
Once you can reliably get a PBS job with the right number of nodes, try
lamboot again; I'm guessing that LAM will do the Right Thing (i.e.,
running "lamnodes" after lamboot will show all the nodes in your job).
If it doesn't, post back to the LAM list, but please include a direct
cut-n-paste from your shell output showing all the steps you took that
results in lamboot not using all the nodes in your job. Thanks!
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
