Bug #44684
closed
pgs entering premerge state that still need backfill
Added by Sage Weil about 4 years ago.
Updated about 4 years ago.
Description
root@reesi002:~# ceph pg ls premerge
PG OBJECTS DEGRADED MISPLACED UNFOUND BYTES OMAP_BYTES* OMAP_KEYS* LOG STATE SINCE VERSION REPORTED UP ACTING SCRUB_STAMP DEEP_SCRUB_STAMP
0.661 12341 0 9421 0 17993553727 0 0 2194 remapped+premerge+backfilling+peered 3m 2255655'4928483 2255789:7875245 [36,135,91]p36 [90,135,91]p90 2020-03-19T19:22:00.122013+0000 2020-03-18T13:07:08.691898+0000
0.e61 12061 0 0 0 17610470273 0 0 898 clean+premerge+peered 3m 2255655'4977032 2255788:8541081 [36,135,91]p36 [36,135,91]p36 2020-03-19T08:28:01.806698+0000 2020-03-16T08:43:25.896120+0000
root@reesi002:~# ceph version
ceph version 15.1.1-155-g3ed3f34 (3ed3f34b9801edf2529295b008b27adee1965640) octopus (rc)
it's the mgr's fault..
2020-03-19T21:12:44.275+0000 7fdbf8aa3700 20 mgr.server adjust_pgs
2020-03-19T21:12:44.275+0000 7fdbf8aa3700 10 mgr.server operator() creating_or_unknown 0 max_creating 1024 left 1024
2020-03-19T21:12:44.275+0000 7fdbf8aa3700 20 mgr.server operator() misplaced_ratio 0 degraded_ratio 0 inactive_pgs_ratio 0 unknown_pgs_ratio 0; target_max_misplaced_ratio 0.06
2020-03-19T21:12:44.275+0000 7fdbf8aa3700 20 mgr.server operator() pool 0 pg_num 3680 target 1024
2020-03-19T21:12:44.275+0000 7fdbf8aa3700 10 mgr.server operator() pool 0 pg_num_target 1024 pg_num 3680 - merge target 0.65f has upmap
then 2 seconds later, before pg stats refresh,
2020-03-19T21:12:46.279+0000 7fdbf8aa3700 20 mgr.server adjust_pgs
2020-03-19T21:12:46.279+0000 7fdbf8aa3700 10 mgr.server operator() creating_or_unknown 0 max_creating 1024 left 1024
2020-03-19T21:12:46.279+0000 7fdbf8aa3700 20 mgr.server operator() misplaced_ratio 0 degraded_ratio 0 inactive_pgs_ratio 0.000221582 unknown_pgs_ratio 0; target_max_misplaced_ratio 0.06
2020-03-19T21:12:46.279+0000 7fdbf8aa3700 20 mgr.server operator() pool 0 pg_num 3680 target 1024
2020-03-19T21:12:46.279+0000 7fdbf8aa3700 10 mgr.server operator() pool 0 pg_num_target 1024 pg_num 3680 -> 3679 (merging 0.e5f and 0.65f)
- Status changed from In Progress to Pending Backport
- Backport set to nautilus
- Copied to Backport #44711: nautilus: pgs entering premerge state that still need backfill added
- Pull request ID set to 34067
- Status changed from Pending Backport to Resolved
While running with --resolve-parent, the script "backport-create-issue" noticed that all backports of this issue are in status "Resolved" or "Rejected".
Also available in: Atom
PDF