NFS-Isilon access from private-astro-cpu

Primary information

Username: briel
Cluster: yggdrasil


Submitting a job array that writes to results in the jobs immediately FAILING with an exit code 0:53, when ran from the private-astro-cpu partition.

I’m still able to access and create files on this filesystem from the login node.
Moreover, it’s also still possible to write to from the debug-cpu`.

These jobs ran from the private-astro-cpu partition used to work fine, but now suddenly fail.

Steps to Reproduce

  1. Create a slurm submission file with the private-astro-cpu partition on and submit it from there.
  2. The submitted run fails.


#SBATCH --array=0
#SBATCH --partition=private-astro-cpu
#SBATCH --ntasks-per-node 1
#SBATCH --time=0-00:05:00
#SBATCH --job-name="mesa_grid_\${SLURM_ARRAY_TASK_ID}"
#SBATCH --output=mesa_grid.%A_%a.out
#SBATCH --mail-type=ALL

echo "test"

Expected Result

I expect the test submit to create a file with mesa_grid_SLURM_ARRAY_TASK_ID.out with ‘test’ written in it.

Actual Result

I get no .out file and the job nearly immediately fails with a 0:53 error code.

It does work again after the nfs-isilon got remounted on the nodes by Remy.