Bug 1505457 - degraded object calcuation is wrong/misleading [NEEDINFO]
Summary: degraded object calcuation is wrong/misleading
Status: CLOSED UPSTREAM
Alias: None
Product: Red Hat Ceph Storage
Classification: Red Hat
Component: RADOS
Version: 3.0
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: rc
: 3.*
Assignee: Josh Durgin
QA Contact: ceph-qe-bugs
Erin Donnelly
URL:
Whiteboard:
Keywords:
Depends On:
Blocks: 1494421
TreeView+ depends on / blocked
 
Reported: 2017-10-23 15:47 UTC by Sage Weil
Modified: 2019-01-23 15:31 UTC (History)
7 users (show)

(edit)
.During a data rebalance of a Ceph cluster, the system might report degraded objects

Under certain circumstances, such as when an OSD is marked out, the number of degraded objects reported during a data rebalance of a Ceph cluster can be too high, in some cases implying a problem where none exists.
Clone Of:
(edit)
Last Closed: 2019-01-23 15:31:00 UTC
ykaul: needinfo? (sweil)


Attachments (Terms of Use)


External Trackers
Tracker ID Priority Status Summary Last Updated
Ceph Project Bug Tracker 21803 None None None 2017-10-23 15:56 UTC

Description Sage Weil 2017-10-23 15:47:50 UTC
Description of problem:

During a data rebalance (no failed OSDs), the system may report degraded objects.  This is alarming and misleading (redundancy is not in fact degraded).  Trivially reproducible on a cluster under any load.

How reproducible:

Easy

Steps to Reproduce:
1. ceph osd pool create foo 64
2. rados -p foo bench 30 write -b 4096 --no-cleanup
3. ceph osd out 0
4. rados -p foo bench 30 write -b 4096 --no-cleanup &
5. watch ceph -s

Actual results:

Shows degraded objects

Expected results:

Only misplaced objects should be shown.

Comment 3 Sage Weil 2017-10-23 15:56:32 UTC
upstream PR: https://github.com/ceph/ceph/pull/18297

Comment 9 Yaniv Kaul 2019-01-23 14:06:08 UTC
(In reply to Sage Weil from comment #3)
> upstream PR: https://github.com/ceph/ceph/pull/18297

This was merged ~1.5 years ago. Can we CLOSE-UPSTREAM this?


Note You need to log in before you can comment on or make changes to this bug.