I have three 3/60's with 24 megs of RAM.

One is dormant, running SunOS 4.1.1 from an ancient 150MB CDC disk when I
need it (haven't needed it in over a year; should find some other use for
this beastie).

Another is dormant because one of the two 600MB disks attached to it has
probably died. Occasionally, I netboot it from the last one, which has run
24x7 for quite some time:

 6:09AM  up 75 days,  9:48, 2 users, load averages: 0.11, 0.09, 0.08

This is chronos.clock.org, running NetBSD 1.1, with the SCSI options
whacked to 3, and only one 1GB Fujitsu M2266 disk. This system gets very
light use - lots of ssh, a little SMTP (a letter every other day?), NTP,
DNS. Occasionally when this guy boots, he gets into the "everything core
dumps" shtick, but I boot him again (and again), until he behaves. Usually
no more than twice or three times.

The aforementioned netboot candidate (quartz.clock.org) does the
"everything core dumps" thing almost invariably. Of course, I've NFS
mounted /usr read-only, where the shared libraries live...

I hate shared libraries precisely because they can cause global mayhem like
this. The interesting question is - if we linked everything static, where
would this error go? Mysteriously dying processes? Random file corruption?

	another data point,

	Erik