On 14-08-15 14:30, Marcin Przyczyna wrote: > Hello, > > this is my first posting to ceph-users mailgroup > and because I am also new to this technology please > be patient with me. > > A description of problem I get stuck follows: > > 3 Monitors are up and running, one of them > is leader, the two are peons. > There is no authentication between the nodes yet. > > Due to my try and error setup I'm afraid I build already > some zombies into my ODS map and > I am not able to get rid of them now. > > First: > "ceph osd tree" reveals: > > # id weight type name up/down reweight > -1 3 root default > -2 1 host arm01 > 1 1 osd.1 DNE > 0 0 osd.0 DNE > -3 1 host arm02 > 2 1 osd.2 DNE > -4 1 host arm03 > 3 1 osd.3 DNE > > Second: > a removal of any of those OSDs seems not be possible anymore: > "ceph osd rm osd.0" produces error "osd.0 does not exist". > Similar error messages occur when I try to remove remaining OSD's > osd.1, osd.2 and osd.3 > You need to remove them from the crushmap as well: $ ceph osd crush remove osd.X Wido > Third: > The OSD processes on all three servers are NOT running. > > Fourth: > OSD' configuration section defines: > [osd] > filestore xattr use omap= true > osd mkfs type = xfs > osd mkfs options xfs = -f > osd mount options xfs = rw,noatime > osd data = /var/lib/ceph/osd/$cluster-$id > osd journal = /var/lib/ceph/osd/journal > osd journal size = 100 > > [osd.1] > host = arm01 > > [osd.2] > host = arm02 > > [osd.3] > host = arm03 > > Fifth: > arch = armhf, > OS = debian 8, > ceph = 0.80.7 > setup = due to absence of convenient setup scripts for armhf > architecture I set up ceph manually. > > Question 1: how can I clean up my OSD map ? > Question 2: What does a column "reweight" wiht value "DNE" mean ? > "Does Not Exist" perhaps ? > > Many thanks in advance, > Regards, > Martin. > _______________________________________________ > 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
 |