Project

General

Profile

Bug #177

unlinked inode during try_to_expire()

Added by Wido den Hollander almost 14 years ago. Updated over 7 years ago.

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

0%

Source:
Tags:
Backport:
Regression:
Severity:
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(FS):
Labels (FS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

After trying to recover from bug #172 my MDS started to crash on their recovery.

Both mds0 and mds1 crashed while trying to recover, with the same trace.

Attached you will find the log and stacktrace. (identical on both MDS's).

mds0_recover_crash_trace.txt View - Debug strace from mds0 and mds1 crash (1.82 KB) Wido den Hollander, 06/04/2010 03:25 AM

mds0_recover_crash_log.txt View - Log entries from crashed MDS (18.6 KB) Wido den Hollander, 06/04/2010 03:25 AM

History

#1 Updated by Sage Weil almost 14 years ago

  • Subject changed from MDS crash on recovery to unlinked inode during try_to_expire()

The problem is that the dirty_inodes list assumes the inodes are either linked or base inodes. That should be the case of any inode on that list.

Finding what caused the unlinked dentry in this case would require the full mds log. Which means just waiting for this to happen again, I think.

#2 Updated by Sage Weil almost 14 years ago

  • Status changed from New to Resolved

Hit this one myself (with full mds logs) and found the bug. Fixed in commit:607ba5dd14e11fe28f1a37cf8a29d4c48d181e0b

#3 Updated by John Spray over 7 years ago

  • Project changed from Ceph to CephFS
  • Category deleted (1)

Bulk updating project=ceph category=mds bugs so that I can remove the MDS category from the Ceph project to avoid confusion.

Also available in: Atom PDF