Project

General

Profile

Bug #52967

premerge pgs may be backfill_wait for a long time

Added by Sage Weil over 2 years ago. Updated over 2 years ago.

Status:
New
Priority:
High
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(RADOS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

             973 active+clean
             146 active+remapped+backfill_wait
             4   active+remapped+backfilling
             2   remapped+premerge+backfill_wait+peered

We should prioritize PGs in premerge state so that they backfill before others, since they are inactive.

Or, perhaps, that should apply to any peered (i.e., inactive) PG?

/a/sage-2021-10-18_15:49:41-orch:rook-wip-sage-testing-2021-10-18-1010-distro-basic-smithi/6448944
(seems to reproduce pretty frequently?)

History

#1 Updated by yite gu over 2 years ago

hi, Sage, What are the conditions for production "premerge+backfill_wait"?

Also available in: Atom PDF