[CentOS] Strange new boot problem with CentOS 4 machine

I have a custom-built PC on which I've been running CentOS 4.x (kept
up to the latest x) for a few years without any problems ... until
this week.

Last week I was out of town on an extended trip, so I shut the machine
down and powered off my UPS, etc.  When I returned on Saturday, I
turned everything back on, and the machine seemed to be booting
normally ... and then it just very abruptly turned itself off, and
refused to power back on.  A short while later I was able to restart
it, but again it shut down during boot.  I *think* all that matters is
that I've waited "long enough," but I was also moving the case around
to open it so there's a slight possibility that something is loose and
being jostled back into connecting (but I really don't think so).

After watching the boot messages carefully and comparing to rc.sysinit
on a different CentOS 4 box, I've concluded that it's shutting down
during loading of the "other" modules (after storage, network, and
audio) parsed from the output of kmodule.  Unfortunately rc.sysinit
doesn't echo anything about what these modules might be.

The only thing that's changed is that I swapped in a new USB  mouse
because the buttons on the old one were getting glitchy, but I tried
leaving that unplugged and the boot still stops at the same place with
the same symptoms.

Does this sound familiar to anyone?  Suggestions on what to do to
either get past this or diagnose it further?