Had two scsi hard drives. One went bad. The bad one had lilo in the MBR. Removed the drive. Then used the Mandrake install cd to boot into rescue mode. Was able to mount all filesystems within rescue mode on remaining hard drive (HD went from sdb to sda). Only glitch was that I could not mount root filesystem as type ext3 -- had to mount as ext2. Modified the lilo.conf file and ran lilo with new parameters. Received an error message when kernel tried to load the root filesystem after rebooting: -- Ext3-fs: invalid journal inode -- mount: error 22 mounting ext3 I then used the rescue disk to boot back in to rescue mode and attempted to run fsck on root filesystem -- /dev/sda1. I get a seg fault when attempting to run fsck on the root filesystem. I remembered one of the messages that I saw mentioned a bad superblock. I mounted the root filesystem to make sure I could. Ran fsck again (I know that's not a good idea to run it on a mounted filesystem). Seg faulted again. Tried to fix the superblock problem by running: -- e2fsck -f -b 8193 on the unmounted file system. Receive the following error message: -- e2fsck: Bad magic number in superblock while trying to open /dev/sda1 -- Superblock could not be read or does not describe a correct ext2 filesystem. Well it's not of type ext2 but of ext3. Tried using a different superblock value (was able to run dump2efs to determine backup superblocks) but still not successful with repairing the filesystem. I ran tune2fs -j /dev/sda1 to make sure the filesystem has the journal inode and it does. I'm at a lost and would like suggestions on possbile solutions. I've thought about copying the root filesystem to another partition, remaking the root filesystem, then copying over back the files. Is there a way to turn the ext3 filesystem back into an ext2? I was hoping that someone could explain what is going on and offer a solution. Thank you Rodney
