[Ocfs2-tools-users] Recovering corrupted files

[Sunil Mushran]

# echo "stat /path/to/file" | debugfs.ocfs2 -n /dev/etherd/e1.0p5

Do you see any extent records? If so, then you should be able to dd the 
data out.

If not, then the search the journals. If you are lucky you could find an
older image of the inode which has the extent records.

# echo "logdump 0" | debugfs.ocfs2 -n /dev/sdX >log0
Do it for all the slots and search for the inode number.

http://oss.oracle.com/projects/ocfs2/dist/documentation/03-disk_format.pdf
You may find this doc relevant.

Chris Poupart wrote:
> Hi Folks,
>
> I have a basic 2 node cluster setup with a shared drive 
> (/dev/etherd/e1.0p5) in the early stages of a test environment.
>
> Unfortunately, this morning a mistake by the power company cut the 
> power for the entire block, and, as this is a test setup, our SAN and 
> 2 nodes were not hooked up to a UPS.
>
> Long story short: 2 files on the shared ocfs2 drive now have a zero 
> file size.
>
> I have the drive unmounted (almost right away), and I have mucked 
> around with debugfs.ocfs2, including attempting to dump one of the 
> files to a local drive, but all I ended up with was a blank file.
>
> I am at a loss on how/if I can repair/retrieve these files.
>
> Any tips would be much appreciated.
>
> -- Chris
>
> _______________________________________________
> Ocfs2-tools-users mailing list
> Ocfs2-tools-users at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-tools-users