Subject: Re: weird/slow disk access
To: Space Case <wormey@eskimo.com>
From: David Brownlee <abs@anim.dreamworks.com>
List: port-i386
Date: 07/04/1998 13:40:35
	All the the oldest ATA (IDE) drives should have automatic bad
	block mapping built into the drive. When the space available
	fills up the bad blocks start 'leaking through'.

	If your drive is showing repeatable bad blocks (the same location
	shows up bad) then I would be very much inclined to check the
	option of a new drive.

	There are ways to keep using the drive - badsect(8) is one.
	If you have a large area of bad sectors you can partition the
	drive around them.

	I have a 2GB drive that chokes badly if I access a certain area,
	currently contained within an 'xemacs-20.3.tgz' file that I've
	just buried away... I really need to get that drive back to Frys
	:)

		David/absolute

                  -=-  "Maybe god will cover up his eyes"  -=-

On Fri, 3 Jul 1998, Space Case wrote:

> I had things going OK, unpacking the source tarballs, when suddenly the
> disk accesses nearly stopped.  The indicator light would come on and I'd
> hear a couple of head movements, then the light would go off and nothing
> happened for ten seconds or so, and repeat ad infinitum.
> 
> After I hit the reset button and it went to fsck the drive, it would go
> merrily along for a while, then exhibit the same behavior.
> 
> I wondered if it might be some kind of disk geometry problem, though the
> drive initially formatted properly.  So I booted the install floppy, mounted
> up the hard disk, and was able to copy the snapshot tarballs and most of the
> source tarballs onto my other system (saving the necessity of downloading
> them again).  When it got part way through the xsrc tarball, the drive
> started acting up again, so now I'm convinced it's a bad sector problem.
> 
> My question is:  What can I do to handle the problem?  (Please keep in mind
> that I'm a newbie concerning PC options, experienced admin otherwise.)
> 
> Here's the specifics:
> 
> Gateway 486DX/33
> Seagate ST32122A IDE
> 
> Error message when transferring xsrc:
> soft error (corrected) reading fsbn 3164800 of 3099264-3099391 (wd0 bn 3456049;
> cn 3428 tn 9 sn 58)
> 
> df output:
>             512-blocks   used       avail
> /dev/wd0a     120798     23170      91588
> /dev/wd0e    3707782    999632    2522760
> 
> disk geom in BIOS: 4096 cyl., 16 hd., 63 sec.
> 
> pfdisk shows partition 4 as ID 165, first cyl 0, last cyl 1023, start
> and length as 63,1032129 sectors
> 
> disklabel wd0 (slightly abbreviated):
> sec/track: 63
> track/cyl: 16
> sec/cyl: 1008
> total sec: 4124736
> rpm: 3600
> 
> 8 partitions:
> #    size   offset    fstype  [fsize bsize cpg]
> a:  124929      63    4.2BSD    1024  8192  16   # (Cyl. 0*-123*)
> b:  166257  124992      swap                     # (Cyl. 124-288*)
> c: 4124673      63    unused       0     0       # (Cyl. 0*-4091*)
> d: 4124736       0    unused       0     0       # (Cyl. 0-4091)
> e: 3833487  291429    4.2BSD    1024  8192  16   # (Cyl. 288*-4091*)
> 
> Thanks,
> ~Steve
> 
> -- 
> Steve Allen - wormey@eskimo.com   http://www.eskimo.com/~wormey/   ICQ 6709819
> 
> Faith is the quality that enables you to eat blackberry jam on a picnic
> without looking to see whether the seeds move.
> 
> Contrary to popular belief, Unix is user friendly.  
> It just happens to be selective about who it makes friends with.
> 	-Kyle Hearn  <kyle@intex.net>
> 
> Weiner's Law of Libraries:
> 	There are no answers, only cross references.
>