Sadly one of the reproduction steps is "have the openstack host's kernel lock up, then hard reboot it", so it's non-trivial to accomplish. To resolve the problem we ended up having to dive into the mysql backend and update the volume statuses manually.
Sadly one of the reproduction steps is "have the openstack host's kernel lock up, then hard reboot it", so it's non-trivial to accomplish. To resolve the problem we ended up having to dive into the mysql backend and update the volume statuses manually.