Path: utzoo!news-server.csri.toronto.edu!rutgers!usc!orion.oac.uci.edu!drector
From: drector@orion.oac.uci.edu (David Rector)
Newsgroups: comp.unix.sysv386
Subject: Re: ISC 2.2 Hangs on disk I/O
Message-ID: <27DD916A.14624@orion.oac.uci.edu>
Date: 13 Mar 91 02:41:46 GMT
References: <24113@hydra.gatech.EDU>
Distribution: usa
Organization: University of California, Irvine
Lines: 59
Bcc: dougp@ico.isc.com, doug@cc.gatech.edu, baxter@slcs.slb.com

This is posted for a friend: Ira Baxter - baxter@slcs.slb.com

------------------------

   I have a Modular Circuit Technology's 25MHz cache 386 system.
   I have a Western Digital WD1006V-SR2 RLL controller with a Segate
   ST-277-1 65MB RLL Harddisk with DOS in the first ~8Meg partion and
   ISC 2.2 UNIX in the rest.  /etc/partions contains an accurate map
   of the bad sectors in the unix partion.

   The problem:
     Some times during disk I/O, the disk light will just stay on solid
     and the system is locked up.  I can not get any response other than
     turning the stupid thing off.

     It happens sometimes during boot up when it is scanning the disk.
     It happens sometimes during large GREP's.
     It just seems to happen!!!
     This is not a panic.  It is just hung.

   Does anyone have any ideas?


I have been living with exactly these symptoms and my WD1006-SRV2 on
my Micronics 386-20 cache motherboard system, with ISC 1.0.6 and ISC
2.0.2.  I actually went (I used to live near them) and beat Western
Digital over the head about it.

All I ever got was "Well, if we can't make it fail here then it
isn't broken."   I never had the guts to take them my system and
let them play with it for two weeks.

I also broadcast (like you) to the net in desperation and got no results.

I can fix the hangup with a simple, hard reset, which is the logical
equivalent of a power off/on cycle, but not CTRL-ALT-DEL.
Forturnately, it only seems to happen to me when my machine is
cold, so I simply don't turn it off anymore.  Consequently I
almost never see the problem (the last time I saw it was 6 months
ago, after I ripped the machine apart to play with its configuration).
This form is pretty livable, but you may not be so lucky.

How old is your controller?  Why don't you send me the numbers written
by WD on it, and I'll compare them to mine; perhaps there is simply a
bad run of controllers.  I personally think it is a bug in the track
buffering logic; something about reading ahead and getting a fault.
But I doubt it will ever get fixed.

If you get any other net feedback, I'd like to hear it.
I'm currently not reading the 386 unix newsgroup; your message
got forwarded to me by another owner of a WD1006, who has had
zero trouble with his on his 486.

baxter@slcs.slb.com

-----------------------
-- 
David L. Rector				drector@orion.oac.uci.edu
Dept. of Math.				U. C. Irvine, Irvine CA 92717