Can you provide context around when this happens - is it immediately after a forced reboot or power loss? This looks like DB corruption we've seen elsewhere with BoltDB.