Actions
Bug #58357
openosd/pg: osd local reservations invalid for PG Deleting
Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:
0%
Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
When the cluster increases or decreases the host OSDs,
After PG is repair to active+clean, the service IO latency is higher than normal,
Set debug_osd=20, the osd log shows that a large number of stray PGs concurrent deleting,
But check osd_ max_ Backfills=1. one osd should have only one stray PG deleting, not multiple,
And ceph daemon osd {x} dump_recovery_reservations, there is no local_server "in progress or waiting",
which is obviously incorrect, indicating that local_reserver for stray PG Deleting
Files
Updated by jianwei zhang over 1 year ago
N/O/P/Q/main
All have the same problems
Actions