Hello! I am attempting to help my brother. I have been speaking with him accross the phone for a long while... and since it is accross the phone I can't copy or paste error messages or give you system specs. However, I do know he is running either SuSe 10.0 or Fedora Core 3. It is a 32bit system. He has two hard drives, his primary (hda) has only one partition, his main partition where Linux is installed. His second drive (hdb) has one partition, this is where Winblows (XP) is installed on an NTFS file system.

He is dual booting with grub.

The problem is when he tries to boot, in any runlevel. When it gets to checking the drives, it marks hda1 as "clean", and then foolishly attempts to check hdb1. This is an NTFS file system, but for some reason it thinks it is an ext3, so it freaks, says the "super-io" is corrupt, and bails to a root "fix it" prompt (with hda as read only).

Now I don't belive the kernel loads it's modules until after the disk check, right? Because he has NTFS installed as a LKM, but that wont help (right?) because it get's loaded AFTER it checks the drives, right?

I had him do a "tune2fs -i 0 -c 0 /dev/hdb1", but it wont do it because it claims it is a corrupted ext3 file system.

How do I force it not to check hdb1? Or is there another problem?

Usually you just set the last field in /etc/fstab to 0 for that mountpoint.

Has he booted up Windows twice since to have it do a file system check to clear any messages in the journal preventing it from being mounted by Linux. Also, it sounds like the entry for it in /etc/fstab may not be correct, if the entry has it as an ntfs file system, there's no reason for Linux to see it as ext3.
Nothing foolish here. If the partition has an entry in /etc/fstab, means it should be mounted at boot-up. It has to be checked before being mounted, if it was mounted when it was corrupt you could possibly loose all data in this partition, what would you think of Linux if that were to happen?. Sounds to me like there may be a small issue broadcasted in the ntfs journal that may only be corrected by Windows, one can also highlight the drive and right click on it to run scandisk on the entire volume via "tools" as Windows will only do an integrity check the parts of the volume that has changed according to the journal since last time it was checked. And make sure the entry in /etc/fstab for the volume is properly configured, you should be able to view/edit the file when presented with a maintenance prompt.

Okay, we fixed it. We did two things at the same time, so I am not sure which it was, but I flushed "/etc/blkid.tab" by doing a:

cat /dev/null > /etc/blkid.tab

(after backing it up of course)

and I also changed his fstab line:

/dev/hdb1 /mnt/Windows/ ntfs auto,user 0 0

to

/dev/hdb1 /mnt/windows/ ntfs default,user 0 0

Thanks for your help guys!

Junior Hacker:

I meant "foolishly" as in trying to check it as an ext3 file system.