Re: Snaptrim_error

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Ah sadly those logs don't look like they have enough debugging to be of much use.

But from what I'm seeing here, I don't think this state should actually hurt anything. It ought to go away the next time you delete a snapshot (maybe only if it has data in those PGs? not sure) and otherwise be ignorable. I've created a ticket at http://tracker.ceph.com/issues/24876
-Greg

On Wed, Jul 11, 2018 at 9:00 AM Flash <flashick@xxxxxxxxx> wrote:
I found only something like this:
2018-07-04 19:26:20.209791 7fc8c8ad2700 -1 log_channel(cluster) log [ERR] : trim_object: Can not trim 11:e4d50fa4:::rbd_data.4d427a238e1f29.0000000000190c9b:6b4 repair needed (no obc)

The full logs in attach. Problem with pg 11.127 started  at ~19:26. Later  11.9 became to error state too, but I don't know, at what time exactly

ср, 11 июл. 2018 г. в 18:20, Gregory Farnum <gfarnum@xxxxxxxxxx>:
On Wed, Jul 11, 2018 at 8:07 AM Flash <flashick@xxxxxxxxx> wrote:
Hi there.

Yesterday I caught that error:
PG_DAMAGED Possible data damage: 2 pgs snaptrim_error
    pg 11.9 is active+clean+snaptrim_error, acting [196,167,32]
    pg 11.127 is active+clean+snaptrim_error, acting [184,138,1]
May it be because the scrub was done when the snapshots were cleaned up?

Hmm, the only way you can get the snaptrim_error state is if the PG gets an error when it tries to trim a particular snapshotted object. And it doesn't get cleared by scrubbing; only when it starts snaptrimming again.

If you have any OSD logs of when this happened, that would be helpful.

And, uh, Sage? Do you know what was supposed to happen here? It's a bit odd as a PG state.
-Greg
 

I tried to restart OSD, then I run deep-scrub and repair, but it didn't solve the problem.

In the documentation the page "Repairing PG inconsistencies" is empty  - http://docs.ceph.com/docs/mimic/rados/operations/pg-repair/,
so I don't know, what else can I do?

Cluster info;
vaersion 12.2.5
25 OSD nodes 
12 OSD per node. The most of them still have filestore as storage backend.

_______________________________________________
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

[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Ceph Dev]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux