? Hello, these days I discovered some strange behavior with snapshot deletes after upgrading from 14.2.22 (nautilus) to 16.2.9 (pacific) The pools are used with librados (no rbd, cephfs or something like that). We create daily one snapshot and delete them after 7 days again. Since the upgrade to pacific the "clones" from the pool corresponding to the snapshot will not be removed nor the space will be freed up. In "rados lssnaps" the snapshot is gone. After I discovered that, I had a look on the old graphs from a small testcluster we upgraded a half a year ago from nautilus to octopus. It seems that "clones" from snapshots which was created with nautilus will not be freed up after removing them with octopus. Snapshots created since the upgrade works fine. It looks like I have now old "clones" which will not be deleted anymore. Is there some way to cleanup the old garbage? And how can I avoid it when upgrading my other clusters, without the option to delete all snapshots prior the update. Manuel? _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |