Project

General

Profile

Bug #37725

mds: stopping MDS with subtrees pinnned cannot finish stopping

Added by Patrick Donnelly almost 2 years ago. Updated 5 months ago.

Status:
Can't reproduce
Priority:
High
Assignee:
Category:
-
Target version:
% Done:

0%

Source:
Development
Tags:
Backport:
octopus,nautilus
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(FS):
MDS
Labels (FS):
Pull request ID:
Crash signature:

Description

Apparently due to checks that prevent export of pinned directories.

This should be reproducible with:

ceph fs set max_mds cephfs 2
mkdir -p foo/bar
setfattr -n ceph.dir.pin -v 1 foo
ceph fs set max_mds cephfs 1
ceph status # should show rank 1 stuck in up:stopping

History

#1 Updated by Patrick Donnelly almost 2 years ago

Actually, this just seems to be really slow when there are lots of subtrees (and large cache without outstanding caps) that are pinned. Zheng, any idea why this would be slow?

#2 Updated by Patrick Donnelly almost 2 years ago

  • Assignee deleted (Zheng Yan)

#3 Updated by Patrick Donnelly over 1 year ago

  • Target version changed from v14.0.0 to v15.0.0

#4 Updated by Patrick Donnelly over 1 year ago

  • Target version deleted (v15.0.0)

#5 Updated by Patrick Donnelly 5 months ago

  • Description updated (diff)
  • Status changed from New to Triaged
  • Assignee set to Xiubo Li
  • Target version set to v16.0.0
  • Backport changed from mimic,luminous to octopus,nautilus

I believe this is fixed already but needs double-checked. It may also be fixed by the ephemeral pinning branch for #41302.

#6 Updated by Xiubo Li 5 months ago

Patrick Donnelly wrote:

I believe this is fixed already but needs double-checked. It may also be fixed by the ephemeral pinning branch for #41302.

Yeah, I have tried for hours and couldn't reproduce it now with the latest upstream code.

#7 Updated by Patrick Donnelly 5 months ago

  • Status changed from Triaged to Can't reproduce

Xiubo Li wrote:

Patrick Donnelly wrote:

I believe this is fixed already but needs double-checked. It may also be fixed by the ephemeral pinning branch for #41302.

Yeah, I have tried for hours and couldn't reproduce it now with the latest upstream code.

Thanks for checking Xiubo. I'll mark this closed.

Also available in: Atom PDF