Project

General

Profile

Actions
Copy link

Bug #37844

open

OSD medium errors do not generate warning or error

Added by Kenneth Waegeman over 5 years ago. Updated about 5 years ago.

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

0%

Source:
Community (user)
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
Ceph - v13.2.1
ceph-qa-suite:
Component(RADOS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

Hi,

I've been seeing inconsistent pgs for a few times past weeks

OSD_SCRUB_ERRORS 1 scrub errors
PG_DAMAGED Possible data damage: 1 pg inconsistent
    pg 5.74d is active+clean+inconsistent, acting [336,113,434,360,368,59,255,163,178,457,283]

With investigating the OSDs part of the pg, I've found that one osd had medium errors in dmesg, and also errors in the ceph-osd log:
2019-01-08 17:56:13.199 7fd3fe1c5700 -1 bluestore(/var/lib/ceph/osd/ceph-457) _do_read bdev-read failed: (5) Input/output error
2019-01-08 18:38:07.751 7fd3fe1c5700 -1 bluestore(/var/lib/ceph/osd/ceph-457) _do_read bdev-read failed: (5) Input/output error

This happened for three disks already the past few weeks.

The problem is there is no warning or error in the ceph health or other ceph commands output that points in that direction.
I would expect having some kind of warning running ceph health detail with failing OSDS.

Thanks!!

Actions
Copy link
 #1

Updated by Greg Farnum about 5 years ago

  • Project changed from Ceph to RADOS
  • Category deleted (OSD)
Actions
Copy link

Also available in: Atom PDF