Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!cornell!uw-beaver!tektronix!tekgen!tekigm2!phils
From: phils@tekigm2.TEK.COM (Philip E Staub)
Newsgroups: comp.sys.amiga
Subject: Re: Bad blocks on hard disks
Message-ID: <3687@tekigm2.TEK.COM>
Date: 24 Oct 88 20:29:36 GMT
References: <4975@louie.udel.EDU>
Reply-To: phils@tekigm2.TEK.COM (Philip E Staub)
Organization: Tektronix, Inc., Beaverton,  OR.
Lines: 42

In article <4975@louie.udel.EDU> AXDRW%ALASKA.BITNET@cunyvm.cuny.edu (Don R. Withey) writes:
[  stuff deleted ]
>say, ah ha! there is where the bad block is...  Then I reliased that I
>had no idea where on the 6 surfaces the bad area was, I just knew the
>cylinder.  So I changed my mount list to stop at cylinder 812.
>
>Now that I have a stop-gap solution to my problem.  How to I go about
>fixing it the correct way???  I have a Seagate ST277N, and a A2090A, I
>looked at the test sheet from Seagate, all it said was PASSED in big letters.
>No error map, like they give you with their ST506 drives...
>
>Any clues anybody???
>        Don
>----------------------------------------------------------------
>Don R Withey                    BITNET: AXDRW@ALASKA.BITNET

This gives me a chance to make a suggestion for some future disk driver(s).

I have a homebrew driver which looks for a port to send error reports to
for either hard or soft errors. If the port is there, it sends the drive,
status, track, head, and sector of the offending block. If not, it just does
its normal error recovery: 3 strikes (retries) and you're out.

Now where is this port? It's created by a program which I run in a separate
window. This program creates the port, then waits for messages (alias: disk
errors), then prints them out. When I get an error, there is no longer any
question where it is. I have visions of a future version which will log
errors to a file (probably in RAD:, since the state of the hard disk driver
may be unknown at this point), but for now this is adequate.

Before anybody asks, this driver is not a replacement for hddisk.device from
Commodore, nor for any other commercial product. It is, as I said, homebrew.

However, I'd like to throw this suggestion out for discussion or whatever.

Regards,
Phil
-- 
------------------------------------------------------------------------------
Phil Staub        
Tektronix, Inc., Vancouver, Washington 98668
phils@tekigm2.MEN.TEK.COM