> Does anyone know of *anything* the 1.2D kernel might have done to 
> cause this strange behavior?  My guess is that this is just a very unhappy 
> coincidence that the machine died shortly after installing a 
> 1.2D kernel.... :-( :-( :-(  (No, I don't blame the 1.2D kernel at all...)

I don't...  One of my multias died not long (~a week) after i updated
it to the 1.2C(?) snapshot, and software bugs were my first thought,
as well.  However, I down-rev'd the software, reinstalled the
firmware, FSL'd the thing, did everything I could to make it go, with
no luck.


> Might it be worth it to install newer firmware? (I'm running 3.5-72 right now, 
> and it's a '94 vintage...)  I'd do that, but I don't want whatever it is 
> that's wrong to fail part-way through the installation, and leave the system 
> completely hosed (like it's not scrambled badly enough already)...

I never had the system die while doing a firmware upgrade (actually,
now that I think about it, maybe i did, once), but even if it does you
can use the Fail-safe loader floppy to fix it up.

While i'd love to say that reinstalling or installing new firmware
will help, I doubt that it will.


> I suspect the poor machine just overheated, and that there is nothing
> I can do to fix it...

That's what my guess is.


The thing that makes/made me skeptical about that notion was the fact
that when mine died, it was a cold day and the window was open, and
the room's fan was on -- the room containing the machine was bloody
cold, to the point that i had to wear a sweatshirt, etc., to be
remotely comfortable.

I seem to recall somebody mentioning that they used
thermistor-controlled fans, so perhaps the thermistor was/is situated
in a place that makes it susceptible to being controlled by the case's
external, rather than internal, temperature, or something...  Could be
that if it's cold, it doesn't make the fan go, and if it doesn't make
the fan go then it stays cold (since it has little/no air flowing over
it), but that seems unlikely.  That would seem to me to be an obvious
design pitfall, that could be easily avoided.



chris