From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Wed,  9 Aug 1995 04:43:07 -0400
From: Nigel Roles ngr@symbionics.co.uk
Subject: help -- 486/33, adaptec scsi, dies on extract of disk2.
Topicbox-Message-UUID: 15044f44-eac8-11e9-9e20-41e7f4b1d025
Message-ID: <19950809084307.WZBl4t428tnOvNabqbURliAe4nkfYg1lLrmvgMkNdcE@z>

> >  /bin/rc
> >  The error was:
> >  rc 30:vdexpand 31:fail |
> >  Press any key to continue...
> 
> I have the same problem (but the exact message varies) on a Dell Optiplex
> 486/66 with Adaptec 1542 controller.
>

This is a DMA overrun on the floppy controller, caused by the disk
cache flushing to the HD as vdexpand goes to the floppy for the next
chunk. For reasons I don't yet understand, the floppy controller 
cannot seize the bus in time. Both the FreeBSD and Linux floppy 
drivers have a "retry forever on DMA overrun" strategy, and change 
comments saying "fixed the DMA overrun bug". Anyone out there know 
why this is such a problem? 

Anyhow, I had exactly the same problem with a similar 1542 based
system, so I bought an IDE drive. I still had the problem; at least
I'm sure I did; now I'm not so sure since two more people are
reporting it with an Adaptec 1542.

Solutions?

1. At least one person I know patched the binary of 9dos so that it 
loads the images from "/n/c:/disk*.vd" not " /n/a:/disk*.vd". He said 
that the floppy still needed to be in the drive, but it read from c:. 
That was on an AMD 386/40 with an Adaptec 1542.

2. Wait till the CD-ROM release!