I boot 2.2pre6 with a new CPU (686MX from 686).
First boot was fine. Went to Win9x to check new CPU,
fine. Reboot 2.2. Unable to mount /var (ext2fs on hdb2)
Ran fsck complained bad-super block. Ran with -b 8193,
fixed errors, mounted fine. Rebooted to check.
Failed to mount /var again. Now fsck, mount, and other
ext2fs utils say not a valid ext2fs, bad magic number.
I tried fsck again with trhe first 10 superblocks,
and get the same results.
Looking at results of 'dd if=/dev/hdb2 bs=1024'
the magic number is the same as a good ext2fs (0xEF53)
at same offset from begining of partition.
Any ideas what is wrong?
How to fix?
-Thomas
The way you present things here it sounds as if Win9x did
something to the superblock. That is possible, such things
have been seen before. But of course something else may
have been wrong, and the intermediate booting of Win9x
might not have had anything to do with these fs problems.
It is a pity that you did e2fsck. (My experience is that
when things are mildly broken, e2fsck fixes everything fine,
but when things are badly broken e2fsck makes things even worse.)
If there are no valuable data on this disk, so that you can play
with it, perhaps you could try to repeat the procedure?
Make a fresh ext2 filesystem on the same partition as before,
Save the partition table and the first 100 blocks or so of this
ext2 partition. Perform the same actions. Report reproducible
results.
My first conjecture would be that nothing is wrong with Win9x
[flames > /dev/null] but that your hardware is not 100% reliable -
maybe this new CPU was overclocked or so. If you get reproducible
problems, do these go away if you use a different CPU?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/