Hi,
did the two replaced OSDs fail at the sime time (before they were
completely drained)? This would most likely mean that both those
failed OSDs contained the other two replicas of this PG. A pg query
should show which OSDs are missing.
You could try with objectstore-tool to export the PG from the
remaining OSD and import it on different OSDs. Or you mark the data as
lost if you don't care about the data and want a healthy state quickly.
Regards,
Eugen
Zitat von "GLE, Vivien" <vivien....@inist.fr>:
Thanks for your help ! This is my new pg stat with no more peering
pgs (after rebooting some OSD)
ceph pg stat ->
498 pgs: 1 active+recovery_unfound+degraded, 3
recovery_unfound+undersized+degraded+remapped+peered, 14
active+clean+scrubbing+deep, 480 active+clean;
36 GiB data, 169 GiB used, 6.2 TiB / 6.4 TiB avail; 8.8 KiB/s rd, 0
B/s wr, 12 op/s; 715/41838 objects degraded (1.709%); 5/13946
objects unfound (0.036%)
ceph pg ls recovery_unfound -> shows that PG are replica 3, tried to
repair but nothing happened
ceph -w ->
osd.1 [ERR] 11.4 has 2 objects unfound and apparently lost
________________________________
De : Frédéric Nass <frederic.n...@clyso.com>
Envoyé : mardi 29 juillet 2025 14:03:37
À : GLE, Vivien
Cc : ceph-users@ceph.io
Objet : Re: [ceph-users] Pgs troubleshooting
Hi Vivien,
Unless you ran 'ceph pg stat' command when peering was occuring, the
37 peering PGs might indicate a temporary peering issue with one or
more OSDs. If that's the case then restarting associated OSDs could
help with the peering or ceph pg. You could list those PGs and
associated OSDs with 'ceph pg ls peering' and trigger peering by
either restarting one common OSD or by using 'ceph pg repeer <pg_id>'.
Regarding the unfound object and its associated backfill_unfound PG,
you could identify this PG with 'ceph pg ls backfill_unfound' and
investigate this PG with 'ceph pg <pg_id> query'. Depending on the
output, you could try running a 'ceph pg repair <pg_id>'. Could you
confirm that this PG is not part of a size=2 pool?
Best regards,
Frédéric.
--
Frédéric Nass
Ceph Ambassador France | Senior Ceph Engineer @ CLYSO
Try our Ceph Analyzer -- https://analyzer.clyso.com/
https://clyso.com | frederic.n...@clyso.com<mailto:frederic.n...@clyso.com>
Le mar. 29 juil. 2025 à 14:19, GLE, Vivien
<vivien....@inist.fr<mailto:vivien....@inist.fr>> a écrit :
Hi,
After replacing 2 OSD (data corruption), this is the stats of my
testing ceph cluster
ceph pg stat
498 pgs: 37 peering, 1 active+remapped+backfilling, 1
active+clean+remapped, 1 active+recovery_wait+undersized+remapped, 1
backfill_unfound+undersized+degraded+remapped+peered, 1
remapped+peering, 12 active+clean+scrubbing+deep, 1
active+undersized, 442 active+clean, 1
active+recovering+undersized+remapped
34 GiB data, 175 GiB used, 6.2 TiB / 6.4 TiB avail; 1.7 KiB/s rd, 1
op/s; 31/39768 objects degraded (0.078%); 6/39768 objects misplaced
(0.015%); 1/13256 objects unfound (0.008%)
ceph osd stat
7 osds: 7 up (since 20h), 7 in (since 20h); epoch: e427538; 4 remapped pgs
Anyone had an idea of where to start to get a healthy cluster ?
Thanks !
Vivien
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io<mailto:ceph-users@ceph.io>
To unsubscribe send an email to
ceph-users-le...@ceph.io<mailto:ceph-users-le...@ceph.io>
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
