pgs inconsistent

S

Senin

Member
Jan 8, 2023
30
7
13
Hi!
3 node cluster, CEPH 17.2.8
after one node had power loss cluster was rebalancing for a while and then 1 pg became active+clean+inconsistent
I've tried to perform deep scrub with ceph pg deep-scrub and after it was finished with no result - ceph pg repair (in progress now)
I read some threads and looks like it doesn't help usually.
What are my options?
The output of the command always shows that OSD.3 was instructed to perform the action, so what if I out this OSD, wait for rebalance and the recreate it?

Best regards, Alex
 
I found that problemtic OSDs were 8 and 10 (both on failed node)
[ERR] : 2.44 shard 8 soid 2:23128f04:::rbd_data.5610ebf5f7019e.0000000000000294:head : candidate had a read error
[ERR] : 2.72 shard 10 soid 2:4fa217a5:::rbd_data.6fabed449625ad.0000000000002183:head : candidate had a read error

and was ready to delete problematic objects wth ceph-objectstore-tool

but fortunately ceph pg repair did the job
 
Last edited:
You must log in or register to reply here.
Top Bottom
Back