Subject: Re: Hard Drive troubles
To: Armen Babikyan <armenb@moof.ai.mit.edu>
From: Dr. Bill Studenmund <wrstuden@loki.stanford.edu>
List: port-mac68k
Date: 10/24/1998 15:02:44
On Sat, 24 Oct 1998, Armen Babikyan wrote:

> Hello,
> 
> I haven't been able to run NetBSD on one of my systems for a few months
> because of an annoying hard drive error. It apparently has some bad sectors.
> when i run fsck -f, i see:
> 
> # fsck -f
> *** /dev/rsd1a
> *** File system is already clean
> *** Last Mounted on /
> *** Phase 1 - Check Blocks and Sizes
> sd1(esp0:2:0): medium error, info = 1417503 (decimal), data = 00 00 00 00
> aa 00 00 00 00 00
> 
> CANNOT READ: BLK 1351872
> CONTINUE? [yn] y
> 
> sd1(esp0:2:0): medium error, info = 1417503 (decimal), data = 00 00 00 00
> aa 00 00 00 00 00
> sd1(esp0:2:0): medium error, info = 1417504 (decimal), data = 00 00 00 00
> aa 00 00 00 00 00
> THE FOLLOWING DISK SECTORS COULD NOT BE READ: 1351875, 1351876,
> *** Phase 2 - Check Pathnames
> *** Phase 3 - Check Connectivity
> *** Phase 4 - Check Reference Counts
> *** Phase 5 - Check Cyl groups
> 25933 files, 492019 used, 168195 free (7459 frags, 20092 blocks, 1.1%
> fragmentation)
> #
> 
> I find it odd that block 1351872 is not listed in "the following disk
> sectors can not be read" part. why not?

Does partition a happen to start at 65629? The fsck errors are within the
partition, while the drive errors aretalking about on-disk sectors. The
difference between them is the start of the partition.

> Is this drive gone for good or should reformatting it do something (like
> eliminate this annoying message)? i thought the drive was supposed to
> automatically map the sector out somehow...

So did I. Either you might have 1) somehow turned off remapping, 2) run
out of remap sectors, or 3) not done the steps needed to remap (I think
remapping only happens on write, but I might be wrong. If so, then if you
only read, you don't remap).

> the drive is a Quantum Lightning 730S 240C according to syslog. it's a
> 730MB drive.
> 
> is there a way to tell the kernel there's a dummy file there, never move
> the file or let anything write there, etc? i wouldn't mind losing a MB of
> diskspace if it meant the disappearance of these problems.
> 
> where on the disk is this bad sector? near the beginning? near the end? or
> how do i find out? if it's in my swap, that could be very bad.

It's somewhere in your a partition. How does 1351875 compare with the
number of blocks in the partition?

Take care,

Bill