Hi all, We have met the following problem: we deployed our env with ceph as volume backend, we boot an instance and attach a ceph volume to this instance, when our nova-compute is down and we delete this instance, it will go local_delete and the ceph volume we attached to this instance will change to "available" status in cinder but when we try to delete it, error will happen, so we will have an "available" volume but can't be attached or delete. We also tested with iscii volumes and it seems fine. I reported a bug about this: https://bugs.launchpad.net/nova/+bug/1672624 Thanks, Kevin Zheng -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20170314/1f4181f6/attachment.html>