[UFO Chicago] the story so far

[Peter A. Peterson II]

Ok, here's what I know so far.

Saturday, my roomate William and I were preparing to host a Quake I LAN
party at my apartment. In order to do so, we wanted to test the various
circuits in the apartment, in the way of finding out how many individual
circuits did we have, were they breakers or fuses, etc. So I shut flynn
down and unplugged it and we commenced to figure out those answers.
Approximately half an hour later when I turned everything back on,
something was wrong. Flynn was running insanely slowly. Like, 286-style
slowly. And Linux doesn't even run on a 286!

I checked the ram, I pulled individual processors (flynn is a dual box),
I booted from floppy -- no dice. I put one of flynn's processors in a
single ppro board that I have along with the scsi and nic, but that
machine crashed intermittently (which may have been a problem of it
trying to run an SMP kernel on a uniprocessor system; some mobos can
handle that, some can't; additionally, while it crashed, it did *not*
act the same way that flynn did). At this point, I thought I had tested
both CPUs and the dual mobo as well as I could, and the show had to go
on.

SO, I ran to work and grabbed two-bit, my workstation with also has
debian and quake I installed. I set up the dhcp server and brought it
home, setting up a small local lan with no natting or outgoing
connection, and we played late into the night.

It was sunday that I tried testing the components again, to no luck.
Finally, I tried putting the scsi and nic into two-bit and booting --
immediate success. So I knew at least that flynn's drives were good, and
that those pieces of hardware were still solid. I replaced two-bit with
a different "brain" -- just a generic workstation from work and returned
two-bit to work. Last night I compiled a uniprocessor kernel and intend
to try booting again on the uniprocessor ppro mobo I have. I *hope* it's
the CPUs and not the motherboard -- pricewatch doesn't even seem to have
any dual ppro boards available any longer, even though I could replace
my CPUs for $80 if I had to. Nick Moffitt suggested that maybe the
onboard caches are borken, so I'll try disabling them.

I'm also now looking into what kind of warranty I had on the motherboard
-- I wonder what they'll do if it's a year warranty but they're out of
stock?

If anyone has any suggestions or ideas, I'm keen to hear them, and I'll
keep you posted. I can use the current configuration as long as I need
to, so things should be basically back to normal, at least on your end.

pedro

-- 
----------------------------------------------------------------------
Peter A. Peterson II, CEO Users of Free Operating Systems, Chicago USA
http://ufo.chicago.il.us -- Free Software Rules -- Proprietary Drools!