superblock blows me out of the water
 

Hi,

I just put a new big disk in my box with MDK 8.2 on it. I partitioned it with 2 * 40GB ish partitions, /usr and /var. Then I rebooted.

It failed on hdb with something about a corrupt superblock. I tried "e2fsck -8 8193 /dev/hdb" that didn't work. I ended up doing "mke2fs -S /dev/hdb1" and "mke2fs -S /dev/hdb5" as I knew these drives were empty anyway.

But now it has been recovering inodes and stuff for about 4 hours and has stalled about 3 times and I have to reboot again. After it has stalled it tries to boot after I hit Ctrl-D and half the services don't start. I don't get it, it is only the new drive (hdb) that is corrupt why don't the services start on hda?

I only want a big database using symbolic links from the mysql directory on hda to exist on hdb. How do I stop the configuration from thinking it has /usr/<all my stuff> and/var/<all my stuff> on hdb instead of hda where I want it to stay?

I hope it recovers OK or I am really in for it with a deadline approaching

Regards

David

how many inodes?
 

Related to my original post. As my disk is recovered (not sure what program is doing this e2fsck?) I note that its fixing illegal blocks in inode nnnnn where nnnnn is currently up to 1680000. How many inodes are there on a 80GB drive with 4K block size? If I know this I might get an idea when it will finish. Since I don't know what an inode is and it's relationship to a block it makes it hard to know khow long the recoverey will take.

Thanks

David

Deep trouble
 

Now the fsck has finished 'fixing' my file system on /dev/hdb but my system fails to boot properly. Half the services in the init fail and I end up with lots of commands not found etc.

I don't understand this at all as it is just my hdb drive that is screwed. All my /usr /var /etc are on /hda drive.

How do i get the system to pretend hdb doesn't exist and get it to boot like it did befoe adding the partitions on hdb (which fail to mount)?