what happens if i run a

helm upgrade

while jobs are running? im assuming configuration gets applied in a safe way?

Running jobs should be unaffected- they don’t pick up configmap/secret changes

This doesn't necessarily apply to dagster jobs since we set

restartPolicy: Never

, but useful to know in general for troubleshooting: • if the configmap/secret is mounted as a volume, the file on the volume is updated, but if the job only reads it once at startup, it shouldn't affect the running job. • if the configmap/secret is mounted as an environment variable, the environment variable will not change • caveat for both: if the job restarts(job spec

restartPolicy

is not

Never

) or gets rescheduled onto another node it will probably pick up the change. It is very unlikely that a job with

restartPolicy: Never

is rescheduled, but kubernetes cannot guarantee this:

Note that even if you specify

.spec.parallelism = 1

and

.spec.completions = 1

and

.spec.template.spec.restartPolicy = "Never"

, the same program may sometimes be started twice.

awesome, thanks @ba