Constant Kernel Panics, Corupted file system reboots over and over

[pseudonym]

Originally Posted by DrJ

The error rate seems high, but it has not changed much since I forced the bad-sector relocation (which differs from a low-level format). That is what you keep your eye on.

How did you do that? Here is the newfs -N /dev/ad8s1d info

Reduced frags per cylinder group from 94088 to 94080 to enlarge last cyl group
/dev/ad8s1d: 715402.3MB (1465144000 sectors) block size 16384, fragment size 2048
        using 3894 cylinder groups of 183.75MB, 11760 blks, 23552 inodes.
super-block backups (for fsck -b #) at:

I should be able to mark that block, or several blocks bad.. just not sure how to do it =(.

[DrJ]

I did it by writing zeros to the entire disk using "dd". It has been a long time, but see the man page for the syntax I have forgotten. Use "if" to be the zero, and "of" to be your disk. There is some info on the smartctl pages somewhere, IIRC. You can even find my posting in the forums, if you like.

[pseudonym]

cool =). thanks! That helps alot. what about using badsect to mark the secotor bad? I was reading something in the handbook about bad144.. but it isn't exactly the clearest section of the book =).

[DrJ]

Originally Posted by pseudonym

what about using badsect to mark the secotor bad?

Sorry, I've not tried it. FWIW, when I did this last, FreeBSD was on 5.2 or a 6 beta -- namely, back in the old days. Back then there were no good tools to locate and mark bad sectors. That may well have changed, since this is one thing that comes up if you use a computer disk. I also recall that there were differences between SCSI and IDE drives, though the forced remapping I suggested should work for either.

If you do use that, you start from ground zero with a decent drive (all other things willing).

[Carpetsmoker]

Newer drives (I guess about 2000 and later) already have spare sectors, if a sector goes bad, it uses a spare sector.

So if bad sectors show up, this means there are already a number of bad sectors, and the number will probably increase exponentially.

By the way, there are a number of disk checking tools on the Ultimate boot CD", which may be able to tell you a bit more about the status/health of your disk

This is a 750GB drive, which is probably rather new ... Most drives come with 4 or 5 years warranty, did you check if you still have warranty?

Originally Posted by DrJ

Sorry, I've not tried it. FWIW, when I did this last, FreeBSD was on 5.2 or a 6 beta -- namely, back in the old days. Back then there were no good tools to locate and mark bad sectors.

Hm?

badsect(8)

HISTORY
The badsect utility appeared in 4.1BSD.