> > > Yesterday night, I made my life miserable by accidentally issuing
> > > the command
> > > mkfs -c /dev/hdb
> > I am afraid you are in serious trouble. What you can do is to try
> > to salvage text from the disk. Try "strings /dev/hdb" and let's
> > hope you can salvage some of the contents of the text files.
> Sad as this case is, I still would be interested to hear if anything
> at all was left on the disk. The '-c' option checks the device for
> bad blocks, so I'd expect it to overwrite the whole disk surface
> with some test patterns before making a filesystem.
I just created a 10 MB file, made a filesystem on it, mounted, and copied
some text to it. Next, I unmounted the filesystem image, made a
"mkfs -c filename", and I still could see the text strings in the file.
the -c switch makes for me only a read-only test.
mke2fs 1.14, 9-Jan-1999 for EXT2 FS 0.5b, 95/08/09
a is not a block special device.
Proceed anyway? (y,n) y
[snip]
Checking for bad blocks (read-only test): done
So, there is still some chance to recover some text fragments from the
screwed disk although I have to acknowledge it is horrible at best.
If you (the one who suffered this accident) have a bigger disk (and
enough free space on it) copy the disk's content into a file on the bigger
drive (dd if=/dev/hdb of=/path/to/huge/space) and try to open it
in a suitable text editor and throw out the unneded stuff. *NEVER* ever
work on the original (formatted) disk but always on a copy. In this
way, there are some chances you can get some things recovered, but it
will be mostly text files. You can try to do this method over and over
and try to salvage a lot of pieces of data. Again, *DO NOT WORK ON THE
ORIGINAL DISK*.
Vilmos