On Thu, 21 Sep 2000, Jeff Hogg wrote:
>
> -----Original Message-----
> From: Ward William E PHDN <[EMAIL PROTECTED]>
> To: Redhat-List (E-mail) <[EMAIL PROTECTED]>
> Date: Thursday, September 21, 2000 3:38 PM
> Subject: Restoring Partition Information
>
>
> >Guys, I need help. Last Friday, one of my wife's hard drives went
> >down in her machine. The drive is defective, but I can still detect
> >that it's there. I need to get the information off the drive, or
> >at least what is salvageable, as she hasn't made a backup in months.
> >
> >This seems like a natural for Linux; I've taken her drive, and inserted
> >it into my machine as hdb. However, the partition information is wiped.
> >The drive was partitioned as FAT32; I'm aware that there are techniques
> >that could be used (for example, I could look at the raw disk and
> >locate text files) but none of the techniques seem to be enough.
as i recall, the original poster suggested the partition table was
toast. i salvaged a drive like this years ago, when i was still
young, stupid and didn't know how hard it should be.
it was on a SCO box, and all i did was write a small C program
that read blocks off of a device and looked for the superblock
"fingerprint" (can't remember exactly what that was, but i think
i looked for the boot block for each partition, and assumed the
superblock started immediately thereafter, which it did.)
this requires nothing more than using the "read" C library call,
and figuring out what is distinctive about an ext2fs filesystem.
using this trick, i did in fact locate the superblocks, hence the
filesystems, and just ran fdisk to replace the partition table.
voila.
so, who wants to suggest how to identify the filesystem? note that
this should work on windows boxes as well, again given some
distinct pattern that identifies the start of a partition.
rday
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
