Computer Science

When you submit your job, the scheduler puts it in a queue. Your job will wait in the queue until it is next in line, and free resources on the compute nodes become available. The cluster’s scheduling policies determine which job is “next in line.” After a job is begun, it runs until it either finishes (with or without errors) or reaches its time limit, at which point the scheduler terminates it.

Use the squeue command to list jobs in queue. The REASON column of the squeue output gives you an idea why your job is not running.

If you see the following error when you submit a job:

sbatch: error: Batch job submission failed: 
Job violates accounting/QOS policy 
(job submit limit, user’s size and/or time limits)

that implies you have hit the maximum number of allowed jobs that can be submitted in a specific partition.

Slurm captures the return value of the batch script/last command and reports this value as the completion status of the job/job step. Slurm indicates status FAILED if the value captured is non-zero.

Make sure you add the following (depending on which shell you use) in your .bashrc or .cshrc file:

BASH

source /etc/profile.d/modules.sh

TCSH

source /etc/profile.d/modules.csh