I see into Ceph and after running command
$ ceph status
I meet text "mds cluster is degraded". I try to seek it at official Ceph docs, but nothing find.
If I understand correctly, MDS needed for mounting and presenting ceph's data as regular FS.
Can someone talk about conditions of appearance for this type of error?
Asked
Active
Viewed 1,574 times
2
Lazy Starwing
- 21
- 3
-
Generically (not ceph-specific), "cluster is degraded" usually means one of your peer-servers is slow to respond, not responding at all, or inconsistent and cannot be synchronized. What do your /var/log/ceph/*mds* logfiles have to say about it? – Andrew Domaszek Dec 19 '14 at 15:16
-
Something like `2014-12-18 16:41:17.885417 7f222eb4a700 0 -- 10.0.0.1:6800/1602 >> 10.0.0.254:0/2864344358 pipe(0x1ba7b200 sd=34 :6800 s=2 pgs=2 cs=1 l=0 c= 0x3b8caa0).fault with nothing to send, going to standby` – Lazy Starwing Dec 19 '14 at 16:29