e2fsck -f /dev/hda3

prints:

group descriptors look bad... trying backup blocks...
resize inode not valid. Recreate<y>?

Should i recreate????

You’re getting error messages that I don’t usually see.

Something is amiss with hda3 with respect to an ext3 filesystem, so you don’t have much to loose by saying “y”. It doesn’t appear to be functional now and may not be after “y”.

Well, if I can save my hard drive by taking it to a local computer shop for a hard drive recovery, will hitting <y> wreck any chances I have for saving what's on it??

I hit <n> and it printed:

Pass 1: checking inodes, blocks, and sizes
Inode 7, i_blocks is 3074, should be 3044. Fix<y>?

I hit <n> to each consequtive question and it told me that the:

filesystem still has errors

I tried it all over again and hit <y> for all questions.
it told me that the file system was modified.

It will only alter what is in hda3, so the NTFS stuff will be safe.

The worst thing that should happen is that hda3 will still be inaccessible, which is what I think is going to happen.

It’s always remotely possible that you put /boot in the LVM, where grub cannot see it, and that hda3 was never formatted.

thing is, i've been using fedora on this computer problem free for as long as fedora 4 has been out. It just all of a sudden has problems.

Grub had read the /boot files fine for months now???

Actually, having seen your post #19, it may be recoverable. So answer “y” to the questions. The LVM stuff will still be there.

I’m going to be off-line for several hours, so other readers of this thread should feel free to offer suggestions and encouragement.

Thankyou WhatsHisName

I don't know what happened, but it booted!!!!!!!

THanks thakns thanks thanks

Just do this:

when the system boots and shows grub,
1. type your "root (hdx,x) " command first
2. then type chainloader +1

This is because there is another grub at the partition where it is installed probably.
Tell me if it works. I load the same way here. If u have grub on a seperate partition instead of MBR, it happens

You can also use <tab> to autocomplete grub entries if you want

At some point in the near future, you should backup /boot and run e2fsck on hda3 from rescue mode without mounting the installation.

For example, you could create /root/boot_backup and use “cp -a ...” to create an archival backup of /boot there.

Then, if the hda3 ext3 filesystem could not be fixed using e2fsck, you could reformat it using “mke2fs -j ...” and then use “cp -a ...” to restore the backup copy to /boot.

If you reformatted hda3 as ext3, you would need to either create 2 mount points and then mount hda3 in one and “/” (i.e., the logical volume) in the other so that you could copy from one to the other OR you could reassemble the entire FC4 setup like this:

mkdir /mnt/sysimage
mount /dev/VolGroup00/LogVol00 /mnt/sysimage
mount /dev/hda3 /mnt/sysimage/boot
chroot /mnt/sysimage
cp -a /root/boot_backup/* /boot

If you had other logical volumes for /home, /var, etc., then you should also mount them in a similar manner before using chroot. And the mount point doesn’t have to be named sysimage. It can be anything you want to call it.

Thanks WhatsHisName

Except that I really understood none of that. If you or someone else could re-explain that to me in really simple newbie language, I'd be so grateful.

Hilary

I had a similar problem while installing FC4. Grub apparently didn't fully install. Most of the files were missing in the grub directory. I had Redhat 9 installed on a different computer, so I copied the grub directory to a thumb drive, then copied them back the the FC4 computer using a livecd. After editing grub.conf, I did get it to start loading the image file, without success. I did do lots of things before I figured out that this was the problem. Now I plan to reinstall FC4, and copy the files from my thumb drive afterwards. hope this helps