On 06/12/11 12:36 AM, David Riley wrote:
On Dec 5, 2011, at 4:05 AM, Toby Thain wrote:
I am sure many on the list will painfully
recognise this scenario:
Some of us remember it profitably. :-) We had a steel mill as a
a year or two ago that needed to move from a Multibus system for which
they could no longer easily obtain replacement parts. We went with a
cPCI system running a Core2 from National Instruments (I wasn't on this
project, but they seemed pretty pleased with it aside from an odd race
condition bug in an Ethernet driver which would bring the whole things
to a screeching halt, which is a bad thing for a steel mill stand
Did they know about comparable bugs in the older system? Can you
describe the older system a bit more? What did it run? (I have three
80386 Multibus systems, Australian Labtam brand).
They had some documentation, but we had to reverse engineer some of
as well... it ended up being a very long process, but we produced
several physical binders worth of information that we made sure they
stashed next to the racks so they didn't get lost.
Interestingly enough, our machines were pressed into service before
had completely finished testing them because one of the Multibus
machines died in a rather permanent way while we were testing its
replacement. It worked, though!