When a concourse worker node is rebooted, does it need to be retired first?

We have an issue where concourse worker nodes are filling their disks after reboots. Do we need to purge CONCOURSE_WORK_DIR after a reboot? Should we retire the worker before rebooting?

Retiring the worker probably wouldn’t hurt. The docs suggest this will purge the volumes. I don’t know if the worker would re-register with the cluster after rebooting though.