[ale] Bad Disk Blocks

[transam at cavu.com (Bob's ALE Mail)]

How about:

	dd if=/dev/hda1 of=/dev/null bs=1024k

Replace hda1 with whatever your disk devices are  ("df" will list hose
mounted.)

If there are any bad blocks then you should see error messages.

Writing to a marginally bad block should clear it.

One possibility if you feel lucky is:

	dd if=/dev/hda1 of=/dev/hda1 conv=sync,noerror bs=512
	(ONLY do this if you fully understand it and have backups
	in case something goes wrong such as a power failure.)

This simply reads each block and then writes it.  If it hits a bad
(unreadable) block then it will write NULs to it, which probably will
clear any CRC errors.  Repeating

	dd if=/dev/hda1 of=/dev/null bs=1024k

will test for this.  The first "dd" listed is always safe (unless
you mistype it) and I do it weekly on about 300GB of disk at the office.

Bob Toxen
bob at cavu.com
transam at cavu.com [ALE]
http://www.mindspring.com/~cavu
Fly-By-Day Consulting, Inc.
"Venimus, Vidimus, Dolavimus" (We came, we saw, we hacked)