[PLUG] Help please, linux md raid broken

[Chris Jantzen]

On Thu, Jan 08, 2004 at 03:09:26PM -0800, Steven Raymond wrote:
> >From: Chris Jantzen <chris at maybe.net>
> >(But if /proc/mdstat says the array is just fine, then, yes, e2fsck
> >and such are the way to go. But still be very, very careful.)
> 
> Again, from the rescue disc:
> 
> e2fsck -nv /dev/md2
> 
> e2fsck 1.32 (09-Nov-2002)
> e2fsck: Invalid argument while trying to open /dev/md2

i.e., /dev/md2 couldn't be auto-detected by the kernel, so it's an
invalid block device. As discussed in previous reply.

> Am lost here.  What's the meaning of "alternate" superblock?  Is the 8193 
> parameter a shot in the dark or should I try the command?

UFS-derived filesystems keep multiple copies of the filesystem
superblock (not the array superblock). If the filesystem itself is
hosed (which we don't know yet, due to previously discussed issues),
you can rebuild the filesystem from one of the backups. Technically
not the same, but metaphorically you can think of it as a "copy of the
FAT table".

-- 
chris kb7rnl =->
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.pdxlinux.org/pipermail/plug/attachments/20040108/eb73e4d5/attachment.asc>