The backend being used for all storage is ceph, with different pools for nova, glance, and cinder; with cinder having a separate pool for ssd and hdd. The goal is being able to migrate VM's from HDD backed storage to SSD backed storage without downtime. Migrating volumes that are not attached works as expected; however, when migrating a volume attached to an instance, the migration appears to fail. I can see the new volume created, and then deleted as the old volume remains. This is the log file for nova-compute during the migration http://paste.openstack.org/raw/691729/ -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack/attachments/20180305/83182e8a/attachment.html>