Topic: Can't Recover From Bad Superblock Error

[Rick Evans]

I am running SME version 5.5 and recently had a superblock error after rebooting the server. I am using two identical 40 gig hard drives in a software raid configuration (/dev/hda & /dev/hdc. I am almost certain that I did not have a power failure but that the problem occurred after I initiated a reboot from server manager.

I have searched both this forum and Google for way to rebuild/correct the problem and have tried the following:

At reboot the server boots only to a command prompt. At that point I have performed the following:

e2fsck on all devices including md0, md1, md2

The recover process will not initiate on devices hda1, hdc1, or md2 because of bad superblock errors. I them performed a "mke2fs -n /dev/hda1," as prescribed in "man e2fsck." This gives me a list of the backup superblocks. I tried each on the reported backup superblocks on each of the devices listed above. The errors I get at that point are either a "Bad Magic Number" or it asks if it is a zero length partition.

I performed a "fdisk -l" to verify the partition information and all seems to check out as valid. (I.e. nothing unusual there)

Is there anything else I can try?

I tried just to upgrade straight to SME 5.6, but the upgrade process will not go until I fix the error.

I have been performing a nightly backup via tape, but I have some customizations I don't want to loose/recreate. I also have some MP3's that I don't backup to save tape space.

Thanks for your help/time,
Rick Evans

[chris]

just typing

fsck /

and see if that helps.

[Rick Evans]

I wish it was that easy. I tried that and many other forms (i.e. e2fsck /; fsck.ext2 /; etc.), however all repeat the same error as described in my original post.

My hope here is that there is a way of recreating the /dev/hda1 partition without a backup superblock.

Thanks for your help.

-Rick