*BSD News Article 16140


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!elroy.jpl.nasa.gov!usc!cs.utexas.edu!uunet!ddsw1!not-for-mail
From: karl@genesis.MCS.COM (Karl Denninger)
Newsgroups: comp.os.386bsd.questions
Subject: New NETBSD user wants to know how to do bad block scan
Date: 14 May 1993 23:48:59 -0500
Organization: MCSNet, Chicago, IL
Lines: 36
Distribution: world
Message-ID: <1t1snr$8lf@genesis.MCS.COM>
NNTP-Posting-Host: localhost.mcs.com

Hi folks,

I got the system to load - sort of.

Turns out that old WD hard disk controllers (MFM style) don't work with
NetBSD -- that was what was generating the panic traps during installation.

However, now I have another question.  

I am running an RLL drive now, or trying to. It has defects, and doesn't
know how to automatically forward sectors.

There appears to be no procedure during the disk setup which scans the disk
and reassigns bad sectors.  Unfortunately, one of them appears in the first
30 or so cylinders, and when the kernel is copied there and then the machine 
reboots it will not load due to the resulting disk error.  I have not been 
able to remap this sector successfully, as I can't get the system loaded far 
enough to attempt manual reassignment of the offending block.

Any ideas?  Does NetBSD assume that you have a disk which appears to be
defect-free to the operating system?  It certainly appears that way!  Do
I need to change to another disk controller and/or disk technology?  

If so, that's disappointing.  This would mean that any grown error on the
disk which the drive doesn't automatically forward could be fatal to the
system as a whole.

There's no mention of a procedure to do this in the installation
instructions.

Any help appreciated.

--
Karl Denninger (karl@genesis.MCS.COM) 	| You can never please everyone except
Modem Access: [+1 312 248-0900]		| by bankrupting yourself.
Voice & FAX: [+1 312 248-8649]		| Internet in Chicago; a MCSNET first!