[Linux-cluster] gfs2 filesystem crash with no recovery?

[Bob Peterson]

----- "Douglas O'Neal" <oneal at dbi.udel.edu> wrote:
| > Is there an alternate superblock that I can use to mount the disk to
| 
| > at least get the last couple of days of data off of it?
| >
| Anybody?

Hi Douglas,

This sounds like either a hardware problem or like someone
trashed the file system by overwriting it.  Unfortunately,
GFS and GFS2 only keep one copy of the superblock.  We've
had some technical discussions about keeping alternate copies
but haven't ever done it yet.

The superblock doesn't contain that much vital information and
could be rebuilt pretty easily.  I could write a tool to do it.
I could also add a new function to fsck.gfs2 to do it as well,
which would be good.  Just need time to do it.
But the problem is that if the superblock was destroyed, it's
likely that the following block, which is the first rgrp block,
would also be destroyed.  The fsck.gfs2 program can rebuild
those, too, depending on the amount of damage.

Perhaps you can dd the first 1MB or so of your file system
to a file, bzip2 it and send it to me and I'll have a look
to see what's wrong.

Regards,

Bob Peterson
Red Hat File Systems