IBM / Bridge-Operator

Bridge operator repo
Apache License 2.0
18 stars 5 forks source link

Slurm samples throw error: status code 400 err Unable to parse query #3

Open soerenmetje opened 1 year ago

soerenmetje commented 1 year ago

Slurm samples: hpcjob-sample0_cm.yaml

kubectl logs hpcjob-pod
I1102 23:05:41.415677       1 main.go:306] Slurm Job with name hpcjob does not exist. Submitting new job.
E1102 23:05:41.441868       1 main.go:154] Submitting job not successful - status code 400 err Unable to parse query.
I1102 23:05:41.441907       1 podutils.go:110] Change in ConfigMap, key jobStatus from value  to value SUBMITTED
I1102 23:05:41.441914       1 podutils.go:110] Change in ConfigMap, key startTime from value  to value 2023-11-02T23:05:41Z
I1102 23:05:41.441919       1 podutils.go:110] Change in ConfigMap, key id from value  to value 0
I1102 23:05:41.457918       1 podutils.go:123] ConfigMap updated.
soerenmetje commented 1 year ago

The parsing of the character " in inline jobScript seems to fail in buildBody function ( https://github.com/IBM/Bridge-Operator/blob/56334fa57caf2de28df6ff76df8a6e6232021421/pods/slurm/main.go#L115C7-L115C7 )

soerenmetje commented 1 year ago

This affects bridgejobs with jobdata.scriptLocation: s3 too. Following example script fails to submit:

#!/bin/bash
#SBATCH --job-name=bridgetest
#SBATCH --output=bridgetest.out

echo "hello world"