Project

General

Profile

Bug #13531

latest update osd lost, cluster still stuck on Peering

Added by shawn chen about 3 years ago. Updated almost 3 years ago.

Status:
Resolved
Priority:
Urgent
Assignee:
Category:
-
Target version:
-
Start date:
10/20/2015
Due date:
% Done:

0%

Source:
Community (dev)
Tags:
Backport:
infernalis, hammer
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:

Description

reproduce procedure:
my test env:

three osds: osd.0 osd.1 osd.2
use rados put xxx xxx -p xxx to put data
suppose osd.1 is primary, then kill -9 osd.0 osd.2

use rados to put xxx data again,
kill -9 osd.1
restart osd.0 and osd.2

ceph -s output seeing that pg down state and ceph pg xxx query seeing that
Peering State blocked by 1
then use ceph osd lost 1 --yes-i-really-mean-it
the Peering State still stuck cause blocked by osd.1

Associated revisions

Revision d5b3926e (diff)
Added by shawn chen about 3 years ago

PG: pg down state blocked by osd.x, lost osd.x cannot solve peering stuck.

Fixes #13531

Signed-off-by: Xiaowei Chen <>

History

#2 Updated by Sage Weil about 3 years ago

  • Status changed from New to Verified
  • Source changed from other to Community (dev)
  • Backport set to infernalis, hammer

#3 Updated by Samuel Just about 3 years ago

This needs a test before merge, shouldn't be hard.

#4 Updated by Kefu Chai about 3 years ago

  • Assignee changed from shawn chen to Kefu Chai

#5 Updated by Kefu Chai about 3 years ago

will try to ping shawn for adding a test.

#6 Updated by Sage Weil about 3 years ago

  • Status changed from Verified to Testing

#7 Updated by Sage Weil about 3 years ago

  • Status changed from Testing to Need Test

#8 Updated by Sage Weil almost 3 years ago

fix is merged, but we still need a test.

#9 Updated by Kefu Chai almost 3 years ago

  • Status changed from Need Test to Resolved
  • Assignee changed from Kefu Chai to shawn chen

Also available in: Atom PDF