HawkBoy
2008-02-29 18:46:44 UTC
Hi,
I apologize if this is an obvious question or if it's been asked
before, but I've searched and can't find anything on it.
I have access to a cluster with 50 nodes. 30 of these nodes are
configured one way and the remaining 20 are slightly different. I have
the misfortune of having code that will only run on the first 30.
I would like to be able to tell qsub to run my jobs on *any* of the
first 30 nodes. Currently, the only (partial) solution I've found is
to specify which node the job is to run on (e.g. qsub myjob -l
nodes=node01), but that means that I have to manually assign nodes to
each job (or assign all to one node and sit in the queue for hours).
Is there something like qsub myjob -l nodes="node01 OR node02 OR ..."?
Again, this may be obvious/easy, but I haven't found anything on it
anywhere.
Thanks
I apologize if this is an obvious question or if it's been asked
before, but I've searched and can't find anything on it.
I have access to a cluster with 50 nodes. 30 of these nodes are
configured one way and the remaining 20 are slightly different. I have
the misfortune of having code that will only run on the first 30.
I would like to be able to tell qsub to run my jobs on *any* of the
first 30 nodes. Currently, the only (partial) solution I've found is
to specify which node the job is to run on (e.g. qsub myjob -l
nodes=node01), but that means that I have to manually assign nodes to
each job (or assign all to one node and sit in the queue for hours).
Is there something like qsub myjob -l nodes="node01 OR node02 OR ..."?
Again, this may be obvious/easy, but I haven't found anything on it
anywhere.
Thanks