Re: Oops while rebalancing, now unmountable.

[Chris Mason <chris.mason@oracle.com>]

Excerpts from Shane Shrybman's message of 2010-11-08 15:39:25 -0500:
> On Mon, 2010-11-08 at 12:55 -0500, Chris Mason wrote:
> > Excerpts from Shane Shrybman's message of 2010-11-08 12:10:57 -0500:
> > > Hi,
> > > 
> > > Got an oops last week while rebalancing that seems to have left me with
> > > a corrupted btrfs. Kernel was ~2.6.36 + Transparent hugetlb patchset +
> > > small misc. patchs.
> > 
> > Have you tried the 2.6.36 + the btrfs unstable git tree?  It may be able
> > to help here.
> 
> Ok, I have built and booted the 2.6.36 + the btrfs unstable git tree
> kernel and ran btrfsck with the same results.
> 
> # btrfsck /dev/sdc1
> checksum verify failed on 625055924224 wanted C3DFFE41 found 9F99998
> checksum verify failed on 625055924224 wanted C3DFFE41 found 9F99998
> checksum verify failed on 625055924224 wanted C3DFFE41 found 9F99998
> Segmentation fault

Ok, if you pull down the latest btrfs-progs from git, you'll find a new
option to btrfsck, which tries alternate copies of the super block.

btrfsck -s 1 /dev/sdc1
btrfsck -s 2 /dev/sdc1
btrfsck -s 3 /dev/sdc1

If we're really lucky, one of these will work.

-chris