31 Dec
1999
31 Dec
'99
6:55 p.m.
----- Original Message -----
From: Bruce Lane <kyrrin(a)bluefeathertech.com>
To: Discussion re-collecting of classic computers
<classiccmp(a)u.washington.edu>
Sent: Saturday, 1 January 2000 6:54
Subject: What Y2K glitch?
We have at least five PeeCees active here at any one
time.
About the same at the cafe here.
Of those five, only ONE got confused about the date
rollover. It was a
1994-vintage 486 system running DOS 6.22. I use it as one of my testbed
systems.
Similar experience. I have a P75 with the buggy Award 4.50G bios.
I deliberately didn't load the y2kure.sys on it to see what would happen.
It got terribly confused (2094!) until I rebooted it with Y2KURE.SYS loaded.
It uses a date offset to ensure the O/S gets a real date, despite the buggy
bios
that won't accept dates not in the 90's. Only other way to fix these
suckers is
to replace the bios. (BTW, this is freeware if you have machines with a
'replace
the bios' type bug.) The machine is still in service in the cafe - so much
for the
"you have to replace your computer" sales pitch.....
Only thing it couldn't fix was an XT that has a real problem with the
GETCLOCK.exe
program that only deals in two digits. The o/s doesn't directly
speak to the clock in an xt, it has to use the getclock program.
y2kure loaded the offset driver ok, but the o/s couldn't get the correct
date through the
usual system calls, and getclock bypasses the y2kure driver...
Manual date setting works if you don't mind setting the clock every restart,
but it was a
bit painful for them.
It's a LASER Turbo (10mhz) XT clone with a RTC board & RLL drive if someone
knows
of a Y2K happy version of getclock for it......
I replaced it with a 286 board circa 1987 that only required a pivot driver
for the rollover
to make the customer happy. (long story - not a lot of money) I also had to
remember
all I thought I'd forgotten about formatting & configuring MFM drives in a
pc, since the
RLL xt 'smart card' controller wouldn't work in the 286, (It pinches
interrupt 8, a Bad
Thing to do to an AT) so I resurrected a 1988 40mb Miniscribe and 16 bit MFM
card
to shift the software onto. That plus the y2kure.sys sorted it all out.
As a bonus, I kept his genuine IBM full length CGA card and monitor and gave
him
an ega card and monitor I had laying about. They're happy and so am I.
(they mostly wanted some games for the kids, and I'm very short of stuff
that will run
on a) an XT and b) CGA.)
EVERY other system, including my old 486-based server,
which any industry
"expert" would happily sneer at as "obsolete," handled the flip-over
without so much as an electronic hiccup.
My Novell 3.12 server required that I manually set the date after a reboot,
to roll
the century over in the hardware clock, but the (y2k patched) Novell had no
probs.
The 4.11 server was fine, it has an sntp client, which saved me having to do
the
reboot and manual reset bit, (Oldish P100 board) as it set the century for
me.
Ah, me... all that hype, and the only glitches
I've heard of this evening
were only minor annoyances (like Auckland's airport web page reporting
flight dates of 1900). Methinks we're going to hear about an awful lot of
annoyed paranoid people and survivalists over the next couple weeks.
A lot of people made a lot of money out of Y2K paranoia. Some of them may
have upset people to deal with shortly. So be it.
The 11 year old Vax 6000 mini at the school here happily rolled over to
2000.
Despite the fact that it's running a 1993 version of VMS (6.0) which is
supposedly
not compliant. (It breaks in a couple of obscure places under obscure
conditions -
none of which occur on this system) Everything here was fine. No problems
that
I am aware of anywhere in the country.
Happy New Year, gang! Keep the peace(es). ;-)
Amen.
Cheers
Geoff Roberts
Computer Systems Manager
Saint Mark's College
Port Pirie, South Australia.
Email: geoffrob(a)stmarks.pp.catholic.edu.au
netcafe(a)pirie.mtx.net.au
ICQ #: 1970476