Bug 16538
Summary: | lvm hangs when creating snapshots of the root partition with 2.6.35 | ||
---|---|---|---|
Product: | IO/Storage | Reporter: | François Valenduc (francoisvalenduc) |
Component: | LVM2/DM | Assignee: | Alasdair G Kergon (agk) |
Status: | CLOSED CODE_FIX | ||
Severity: | normal | CC: | agk, akpm, fs_ext4 |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | Subsystem: | ||
Regression: | Yes | Bisected commit-id: | |
Attachments: |
Output of lvmdump
Output of strace when running lvcreate |
Description
François Valenduc
2010-08-07 14:36:00 UTC
Try to supply some diagnostics - e.g. the process backtraces sysrq will show and 'lvmdump' after rebooting. Created attachment 27372 [details]
Output of lvmdump
So here is the output of lvmdump after having restarted the computer.
So, the logical volume is well created but it's not a snapshot. When I try to mount it, I get the following error: mount: unknown filesystem type 'DM_snapshot_cow' zsh: exit 32 LC_ALL='C' mount /dev/gentoo/backup /mnt/backup I noticed that this problem occurs when I take a snapshot of the root partition. I also tried git bisect but it doesn't reveals anything conclusive. The first bad commit is supposed to be the following: f3b99be19ded511a1bf05a148276239d9f13eefa is the first bad commit commit f3b99be19ded511a1bf05a148276239d9f13eefa Author: NeilBrown <neilb@suse.de> Date: Thu Jun 24 13:31:03 2010 +1000 Restore partition detection of newly created md arrays. However, I don't use RAID but only LVM so this is not the real first bad commit. Created attachment 27375 [details]
Output of strace when running lvcreate
Maybe you will find the output of strace interesting. Unfortunately, this is so long that I can only get the end. This is thus the result of
I did something wrong. So, the strace is the result of strace lvcreate -L1G -n backup -s /dev/gentoo/root. Since it completely fills the console, I only have the end of it. I finally find the problematic commit with git-bisect. This is the following one: commit 6b0310fbf087ad6e9e3b8392adca97cd77184084 Author: Eric Sandeen <sandeen@redhat.com> Date: Sun May 16 02:00:00 2010 -0400 ext4: don't return to userspace after freezing the fs with a mutex held If I revert it, the problem doesn't occur. This problem is in fact fixed by the following commit: commit d2aa412a7e270bc679a586dda415344f2b8ad01e Author: Eric Sandeen <sandeen@sandeen.net> Date: Sun Aug 1 17:33:29 2010 -0400 ext4: fix freeze deadlock under IO |