Dear everyone,
during a very risky process or rebuilding my OSDs to relocate their
WAL&DBs onto NVMe SSD partitions, I created a situation where 2 of 3
OSDs of jerasure 2+1 pool had become destroyed, without waiting for
complete rebuild of their respective PGs on newly-created OSDs.
And now I have an incomplete PG, looking like this :
[WRN] PG_AVAILABILITY: Reduced data availability: 1 pg inactive, 1 pg
incomplete
pg 1.90 is incomplete, acting [2,3,5] (reducing pool jerasure21
min_size from 2 may help; search ceph.com/docs for 'incomplete')
PG OBJECTS DEGRADED MISPLACED UNFOUND BYTES OMAP_BYTES*
OMAP_KEYS* LOG STATE SINCE VERSION REPORTED UP ACTING
SCRUB_STAMP DEEP_SCRUB_STAMP
1.90 0 0 0 0 0 0 0 0
incomplete 47m 0'0 223998:104494 [2,3,5]p2 [2,3,5]p2
2020-04-01T13:22:08.983764+0300 2020-03-29T18:36:21.037198+0300
This corrupted some (or maybe all) of my iSCSI RBD targets, and now I
would like to determine what of them will have to be dropped & recreated
so as not to refer to these incomplete objects. Is there a known recipe
or scenario to do that ? Also I would like to know if anything special
needs to be done to the incomplete PG to make it complete and usable again.
Thanks in advance for your help.
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
