Project

General

Profile

Bug #16988

'rbd feature disable' against non-primary image results in hung CLI

Added by Jason Dillaman 9 months ago. Updated about 1 month ago.

Status:
Resolved
Priority:
Low
Target version:
-
Start date:
08/10/2016
Due date:
% Done:

0%

Source:
other
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Release:
Needs Doc:
No

Description

The command will continuously repeat an error about requesting the exclusive lock:

$ rbd --cluster cluster1 --pool mirror feature disable test journaling 
2016-08-10 15:01:29.470815 7f699985d180 -1 WARNING: the following dangerous and experimental features are enabled: *
2016-08-10 15:01:29.471004 7f699985d180 -1 WARNING: the following dangerous and experimental features are enabled: *
2016-08-10 15:01:29.472940 7f699985d180 -1 WARNING: the following dangerous and experimental features are enabled: *
2016-08-10 15:01:29.489846 7f699985d180 -1 librbd: Mirrored image is not the primary, add force option to disable mirroring
2016-08-10 15:01:29.489858 7f699985d180 -1 librbd: error disabling image mirroring: (22) Invalid argument
2016-08-10 15:01:29.493150 7f69727fc700 -1 librbd::ImageWatcher: 0x7f69540008c0 error requesting lock: (30) Read-only file system
2016-08-10 15:01:30.494427 7f69727fc700 -1 librbd::ImageWatcher: 0x7f69540008c0 error requesting lock: (30) Read-only file system
2016-08-10 15:01:31.495998 7f69727fc700 -1 librbd::ImageWatcher: 0x7f69540008c0 error requesting lock: (30) Read-only file system
2016-08-10 15:01:32.497693 7f69727fc700 -1 librbd::ImageWatcher: 0x7f69540008c0 error requesting lock: (30) Read-only file system
2016-08-10 15:01:33.499159 7f69727fc700 -1 librbd::ImageWatcher: 0x7f69540008c0 error requesting lock: (30) Read-only file system
2016-08-10 15:01:34.500463 7f69727fc700 -1 librbd::ImageWatcher: 0x7f69540008c0 error requesting lock: (30) Read-only file system
...

History

#1 Updated by Jason Dillaman 9 months ago

  • Backport set to jewel

#2 Updated by Brett Niver 8 months ago

  • Status changed from New to In Progress
  • Assignee set to Brett Niver

Seems only one line of code missing - at least to avoid the hang, and allows an error message to be displayed.

#3 Updated by Venky Shankar 7 months ago

  • Needs Doc set to No

Jason, I think Mykola's PR https://github.com/ceph/ceph/pull/10980 would automatically fix this, right?

#4 Updated by Venky Shankar 7 months ago

  • Status changed from In Progress to Resolved

Tested with PR#10980, no cli hang was seen:

./bin/rbd --cluster cluster2 -p mirror feature disable r0 journaling
2016-09-27 13:30:57.037843 7fad3f2f4180 -1 WARNING: the following dangerous and experimental features are enabled: *
2016-09-27 13:30:57.037890 7fad3f2f4180 -1 WARNING: the following dangerous and experimental features are enabled: *
2016-09-27 13:30:57.039812 7fad3f2f4180 -1 WARNING: the following dangerous and experimental features are enabled: *
2016-09-27 13:30:57.062617 7fad0f7fe700 -1 librbd::journal::StandardPolicy: local image not promoted
2016-09-27 13:30:57.062625 7fad0f7fe700 -1 librbd::exclusive_lock::AcquireRequest: failed to allocate journal tag: (1) Operation not permitted
2016-09-27 13:30:57.071543 7fad0f7fe700 -1 librbd::ExclusiveLock: failed to acquire exclusive lock:(1) Operation not permitted
rbd: failed to update image features: (30) Read-only file system

#5 Updated by Venky Shankar 7 months ago

  • Status changed from Resolved to In Progress

Moved to resolved too quickly, should be done after PR#10980 is merged.

#6 Updated by Jason Dillaman 4 months ago

  • Assignee changed from Brett Niver to Jason Dillaman

#7 Updated by Jason Dillaman 4 months ago

  • Status changed from In Progress to Need Review

PR: https://github.com/ceph/ceph/pull/12639 (just the "librbd: fail immediately if the exclusive lock cannot be acquired" portion)

#8 Updated by Mykola Golub 4 months ago

  • Status changed from Need Review to Pending Backport

#9 Updated by Jason Dillaman 4 months ago

  • Backport changed from jewel to kraken,jewel

#10 Updated by Jason Dillaman 4 months ago

  • Status changed from Pending Backport to Resolved
  • Backport deleted (kraken,jewel)

On second thought: it will be backported by #18327

#11 Updated by Jason Dillaman about 1 month ago

  • Status changed from Resolved to In Progress

Fix was lost in master branch

#12 Updated by Jason Dillaman about 1 month ago

  • Status changed from In Progress to Resolved

Verified that it still works in master after exclusive lock / managed lock refactor

Also available in: Atom PDF