On Thu, May 29, 2014 at 1:48 AM, Scottix <scottix at gmail.com> wrote: > // Then tried but got error > ceph osd pool delete metadata metadata --yes-i-really-really-mean-it > Error EBUSY: pool 'metadata' is in use by CephFS This is issue #8010. Previously, there was no check for pools being in use before they were deleted, which was kind of dangerous - in firefly we added the check, and this is the consequence. There will soon (https://github.com/ceph/ceph/pull/1852) be a "rm" command that will let a user fully disable the filesystem, so that they can remove the pools. We willl also no longer be creating the filesystem pools by default. > // Tried and it looks like a bug of sort > ceph mds cluster_down > // Still get > mdsmap e78: 0/0/0 up > // Shouldn't it be down? That's telling you that there are "zero MDSs up", rather than saying that something is up. > Do I need to start over and not add the mds to be clean? You're already clean, apart from having a couple of unwanted pools hanging around. If you're worried about resource consumption from those pools, you could create some very-small-pg_num pools and use newfs to switch the FS to those. Cheers, John
 |