Actually, it is. We took the single host getting started out, because nobody would really deploy a distributed system like Ceph for production on single host. The problem is that the default crush rule is set to the host level, not the osd level.
Note, I think ceph-deploy mon create-initial will do the next two steps for you. So those may be redundant. What you need to do though is after you do ceph-deploy new ceph-a1, is you need to add the following to your ceph.conf file: osd crush chooseleaf type = 0 Then, follow the rest of the procedure. On Fri, Jan 31, 2014 at 2:41 PM, Cristian Falcas <[email protected]>wrote: > Hi list, > > I'm trying to play with ceph, but I can't get the machine to reach a > clean state. > > How I did the installation: > > ceph-deploy new ceph-a1 > ceph-deploy install ceph-a1 > ceph-deploy mon create-initial > ceph-deploy mon create ceph-a1 > ceph-deploy gatherkeys ceph-a1 > > ceph-deploy disk zap ceph-a1:vdb ceph-a1:vdc ceph-a1:vdd ceph-a1:vde > ceph-deploy osd prepare ceph-a1:vdb ceph-a1:vdc ceph-a1:vdd ceph-a1:vde > ceph-deploy osd activate ceph-a1:/dev/vdb ceph-a1:/dev/vdc > ceph-a1:/dev/vdd ceph-a1:/dev/vde > > What the status is: > [root@ceph-a1 ~]# ceph health > HEALTH_WARN 49 pgs degraded; 192 pgs stuck unclean > > cceph -w: > 2014-01-31 17:39:44.060937 mon.0 [INF] pgmap v25: 192 pgs: 102 active, > 41 active+remapped, 49 active+degraded; 0 bytes data, 143 MB used, 243 > GB / 243 GB avail > > > Even if I add more disks or play with the crush map settings, I can't > seem to manage to bring the PGs to a clean state. > > Is this expected with one host only? > > Best regards, > Cristian Falcas > _______________________________________________ > ceph-users mailing list > [email protected] > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > -- John Wilkins Senior Technical Writer Intank [email protected] (415) 425-9599 http://inktank.com
_______________________________________________ ceph-users mailing list [email protected] http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
