=46ascinating!

My principal machine "cesium.clock.org" is a Sun 4/330 with 32 MB RAM, and
two 3.5 GB disks (DEC DSP 5350S), running SunOS 4.1.4 (has been up more or
less continuously for a few years since I got it, used even then).

I have also been having "DATA FAULT" panics lately, after a couple of years
of stable operation in an environmentally controlled machine room - no
changes in the hardware since last summer. Since all Suns use parity RAM
(which, one assumes, would catch a bad SIMM, and the kernel would report as
a "parity" or "memory" error), this has to be something else. A friend has
suggested bad SCSI cabling or some kind of disk error (none reported by the
kernel yet), but I can't shake the feeling that my hardware is simply
getting old, and some chip somewhere has gone marginal on me - that is,
it's time to upgrade to something else (I have this nice SPARC classic that
is being prepped now).

If any of you might have a more concrete notion what might be causing these
panics, I'd love to hear it. I have the dmesg output for the last four
crashes, for the curious.

While I recognize that this is not directly a NetBSD issue, I suggest that
since most of us are running on castoff hardware, hardware failure due to
age is a class of problem that we're going to see more of, and if we can
put some insight into the list archives, it may serve those who come after
us that much better.

	thanks,

	Erik Fair <fair@clock.org>