Ugh - need some help. I think the hard drive is failing.

I have tried a few programs which failed. I am currently running clonezilla for about 24 hours. It was running quick with an Esata cable, but slowed to a crawl with tons of buffer, i/o errors ,etc. It is still running, so I will let it be for awhile.

My next venture will be to use ddrescue. I am planning on using:

sudo ddrescue -v -r 3 /dev/sda3 /dev/sdb3 logfile

I know there are going to be a ton of errors, bad sectors and I think the above will cover the skipping of those. Please be as specific as possible

Any other ideas for me? Should I run a Western Digital test/repair on it?

Oh yeah, I can boot up, but runs completely slow so I have been booting from the CD.

Any help for a new user would be appreciated. I was able to copy most of my files off with knoppix, now just looking to do a usable bootable clone.

I don't quite understand why you are trying to clone the system if you have already recovered most of your files. Usually it would be quicker to reinstall than to attempt to completely clone it.

Even if you do succeed in making a bootable copy, it will not be trustworthy, since there may be critical sectors missing (that you don't discover until you do a fsck or actually use a file which has missing sectors of data).

If there are still sections that have not been recovered because clonezilla got bogged down, then by all means use ddrescue with a small retry limit to image the partition; but afterwards simply mount the rescued partition on a working system and grab the files you want, no need to waste time attempting to boot from it. There is no need to copy the partition to another (same sized) partition; just copy it to a large file on another drive which can then be loop mounted. Normally this copy would be the first step, since repeated access to bad sectors at best wastes valuable recovery time and at worst may exacerbate the problems.

If you are certain that the errors are disk errors (and not say a cable fault), then there is little point running a test/repair on it. This would only be worthwhile as a last ditch attempt to recover critical files. The disk itself should be quietly retired.