Project

General

Profile

Actions
Copy link

Bug #53056

open

rbd map fail when has a file named /dev/rbd0 which is the mapping target

Added by yunqing wang over 2 years ago. Updated about 2 years ago.

Status:
Need More Info
Priority:
Normal
Assignee:
Ilya Dryomov
Target version:
-
% Done:

0%

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

Description

This happened when I start a fio to test the /dev/rbd0 when rbd map failed.
The fio create a file named /dev/rbd0. After that, I remapped rbd, and it get success. But /dev/rbd0 is still a file not a device.

It can be simply trigged as

touch /dev/rbd0
rbd map 1

rbd showmapped
id pool namespace image snap device
0 rbd 1 - /dev/rbd0

but /dev/rbd0 looks like rw-r--r- 1 root root 0 Oct 27 11:07 /dev/rbd0

uname -a
Linux node1 5.4.0-81-generic #91~18.04.1-Ubuntu SMP Fri Jul 23 13:36:29 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

Related issues 1 (0 open — 1 closed)

Related to rbd - Bug #49410: krbd: make sure the device node is accessible after the mappingResolvedIlya Dryomov

Actions
Actions
Copy link
 #1

Updated by Sebastian Wagner over 2 years ago

  • Project changed from Ceph to rbd
  • Category deleted (librbd)
Actions
Copy link
 #2

Updated by Ilya Dryomov about 2 years ago

  • Status changed from New to Need More Info
  • Assignee set to Ilya Dryomov

Hi Yunqing,

Which version of rbd CLI tool is in use? You can check with "rbd --version".

I believe this has been addressed in pacific (16.2.*) and recent octopus point releases (starting with 15.2.13).

Actions
Copy link
 #3

Updated by Ilya Dryomov about 2 years ago

  • Related to Bug #49410: krbd: make sure the device node is accessible after the mapping added
Actions
Copy link

Also available in: Atom PDF