On Mon, Mar 13, 2017 at 2:13 PM, Kent Borg <kentborg@xxxxxxxx> wrote: > We have a Cephfs cluster stuck in read-only mode, looks like following the > Disaster Recovery steps, is it a good idea to first make a RADOS snapshot of > the Cephfs pools? Or are there ways that could make matters worse? Don't do this. Pool-level snapshots are incompatible with so-called "self-managed" snapshots where something above rados (cephfs/rbd) is handling snapshots. As the docs suggest, you should always take a backup of your journal. You can even take a full copy of the metadata pool with "rados export" if it's manageably small (metadata often is). John > Thanks, > > -kb > > _______________________________________________ > ceph-users mailing list > ceph-users@xxxxxxxxxxxxxx > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |