[slurm-users] Query about sbatch: Request X nodes in a range?

Thompson, Matt (GSFC-610.1)[SCIENCE SYSTEMS AND APPLICATIONS INC] matthew.thompson at nasa.gov
Wed Jul 14 17:04:45 UTC 2021


A simple user SLURM question. Recently I had a need to request a certain range of nodes on a cluster, and I found SLURM was acting against my assumption, but exactly as the man page says.

Namely, I needed say, 20 nodes on a cluster on the same rack (benchmarking!), but I suspected that some node was down for various reasons. So that if I tried something a la:

  sbatch -N 20 --nodelist=nodeX[001-020]

the job would be stuck pending. My naïve thought was oh, okay, I'll just ask for a bigger swath, like the whole rack (say 50 nodes):

 sbatch -N 20 --nodelist=nodeX[001-050]

and it will give me 20 nodes within that range, and skip the one that's down. But that lead to:

  sbatch: error: invalid number of nodes (-N 50-20)

I then read the man page and found, that nodelist will let you do the "under":

  sbatch -N 20 --nodelist=nodeX[001-010]

but not the "over". Instead I needed to do:

  sbatch -N 20 --nodelist=nodeX[001-011,13-21]

after finding out, for example, node 12 was down.

So, my question is, is there a way to say, "Please give me X nodes inside this specific range of nodes?"

Matt Thompson, SSAI, Ld Scientific Programmer/Analyst
NASA GSFC,    Global Modeling and Assimilation Office
Code 610.1,  8800 Greenbelt Rd,  Greenbelt,  MD 20771
Phone: 301-614-6712                 Fax: 301-614-6246
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20210714/e777214d/attachment-0001.htm>

More information about the slurm-users mailing list