What does your crushmap look like? There's a good chance you're
choosing first hosts, and then OSDs, which means you can't come up with
3 replicas (because there are only two hosts).
Try:
ceph -o my.crush.map osd getcrushmap
crushtool -i my.crush.map --test --output-csv
and then look at the .csv files created in that directory; that
simulates some random object placements, and will let you know which
OSDs the crushmap chose. I bet you'll see that the data pool isn't
replicating to 3 OSDs.
On 05/20/2013 11:51 AM, Nicolas Fernandez wrote:
Hello,
I'm deploying a test cluster on 0.61.2 version between two nodes
(OSD/MDS), and another (MON).
I have a problem making my cluster grow, today i've added an OSD into a
node that was a osd exist. I've made a reweight and add a replica.The
crushmap is up to date but now i'm getting some pgs in stuck
unclean.I've been cheecking tuneables options but that haven't sold the
issue, how can i fix the healthof the cluster?.
My cluster status:
# ceph -s
health HEALTH_WARN 192 pgs degraded; 177 pgs stuck unclean; recovery
10910/32838 degraded (33.224%); clock skew detected on mon.b
monmap e1: 3 mons at
{a=192.168.2.144:6789/0,b=192.168.2.194:6789/0,c=192.168.2.145:6789/0
<http://192.168.2.144:6789/0,b=192.168.2.194:6789/0,c=192.168.2.145:6789/0>},
election epoch 148, quorum 0,1,2 a,b,c
osdmap e576: 3 osds: 3 up, 3 in
pgmap v17715: 576 pgs: 79 active, 305 active+clean, 98
active+degraded, 94 active+clean+degraded; 1837 MB data, 6778 MB used,
440 GB / 446 GB avail; 10910/32838 degraded (33.224%)
mdsmap e136: 1/1/1 up {0=a=up:active}
The replica configuration is:
pool 0 'data' rep size 3 min_size 2 crush_ruleset 0 object_hash rjenkins
pg_num 192 pgp_num 192 last_change 576 owner 0 crash_replay_interval 45
pool 1 'metadata' rep size 2 min_size 1 crush_ruleset 1 object_hash
rjenkins pg_num 192 pgp_num 192 last_change 556 owner 0
pool 2 'rbd' rep size 2 min_size 1 crush_ruleset 2 object_hash rjenkins
pg_num 192 pgp_num 192 last_change 1 owner 0
OSD Tree:
#ceph osd tree
# id weight type name up/down reweight
-1 3 root default
-3 3 rack unknownrack
-2 1 host ceph01
0 1 osd.0 up 1
-4 2 host ceph02
1 1 osd.1 up 1
2 1 osd.2 up 1
Thanks.
_______________________________________________
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
