KRaft currently does not support JBOD. However, it uses the data volumes to store some cluster metadata on it (such as the ID). Strimzi currently allows removing JBOD volumes, so once KRaft supports JBOD, we need to carefully check that we do not have some loophole to screw the whole cluster.

There might be no action for this issue if Kafka solves it on its own somehow - but if nothing else this Jira is a reminder of a possible problem which needs to be checked in the future.

Created by Strimzi#9437