Check error condition for Jacamar jobs to rerun the job in face of node failure
The Jacamar jobs will sometimes fail due to node failure. This can result in the CI job failing, and can generally be easily fixed by check if the failure is due to node failure.
We should detect whether the job failure is due to a node failure (by faithfully catching the NF flag from slurm), and then rerun the job. This will prevent spurious failures in the CI.