Re: BTRFS Kernel Panic
On Thu, Apr 11, 2013 at 03:50:19PM -0500, Alex Robbins wrote:
> So it looks like btrfs really is still experimental.
> I have / and /var on two separate btrfs partitions and I consistently get
> a kernel panic when I run "aptitude update" (see the end). I also managed
> to get the same results by using find and dd to read every file in /var.
> Although my oops output does have RIP: btrfs_num_copies, I do not think that
> I need to use btrfs-zero-log because (1) I am using a 3.2 series kernel, (2)
> my panic callstack does not have a function that begins with "replay_one_",
> and (3) I am able to mount the partitions (and boot) just fine. It is only
> when some process such as aptitude or updatedb (indexing for the locate
> command) tries to read something in the filesystem.
I'd suggest that you firstly:
1) take a full image of the raw devices containing these filesystems
with dd in case of further damage
2) try to backup the filesystems with tar
If you get a backup, you could then reformat the partitions and
restore the backups.
I'd definitely recommend looking at the changelog for newer
versions of the linux kernel to see if this has fixed. It
also might be worth contacting the brtfs developers' mailing
list to see if they have seen this issue before--the BUG alone
might be useful.
You could also try upgrading to a newer kernel e.g. 3.8.x. I've
done this myself due to btrfs issues with older kernels; you might
need to hand-build it though since Debian doesn't yet have it.
[I've previously lost a large raid-1 btrfs filesystem--both copies
completely trashed after a SATA glitch, including the good copy
which was corrupted entirely by btrfs... Currently using it on an
SSH for testing purposes as my rootfs, but don't use it for user
data, which is on ext4 on LVM RAID.]
.''`. Roger Leigh
: :' : Debian GNU/Linux http://people.debian.org/~rleigh/
`. `' schroot and sbuild http://alioth.debian.org/projects/buildd-tools
`- GPG Public Key F33D 281D 470A B443 6756 147C 07B3 C8BC 4083 E800