On Mon, Nov 4, 2019 at 11:04 PM J David <j.david.lists@xxxxxxxxx> wrote: > > OK. Is there possibly a more surgical approach? It's going to take a > really long time to convert the cluster, so we don't want to do > anything global that might cause weirdness if any of the OSD servers > with unconverted OSD's need to be rebooted during the process. FWIW we used a brute-force approach here in the past. A loop that tries up to 10 time to kill all processes that show up in lsof for that disk, unmounts it, and wipes it. ceph-volume solved all of these problems Paul _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |