> >> I'm running NetBSD 1.5.1 on a Pentium 100 with 8 megabytes of ram.
> >it's much probably this machine isn't good. memory is the first thing to
> >check.
>
> This machine has run Windows for 2 years, Linux for approximately 3 years..
> With Linux it had an uptime of about one year and regularly ran
> X11, so did a lot of swapping.
so it's another problem
> 3 weeks. It happened just now, seemingly out of the blue. It is not reproducable
> as far as I can see. (Again, suprisingly similar to the openbsd posting)
> However, the external cache is now disabled.
>
> >> is this just memory exhaustion?
> >i'm using netbsd on 486 25-66Mhz, 8MB RAM+12MB swap and got over 100 days
> >uptime on one machine. this machines are used for NAT, but some other
> >programs are on them.
>
> % swapctl -l
> Device      512-blocks     Used    Avail Capacity  Priority
> /dev/wd0b       133245     2512   130733     2%    0

with netbsd for every 256MB of swap 1MB of real memory is LOCKED (16
bytes structure for each 4KB page) . so 40-50MB is better choice.

> It was running, at the time of crash, rpcbind, nfsd, rpc.bootparams, inetd, rarpd, sshd, cron and syslogd, as well as a fairly full ipnat and ipf ruleset.

what is your kernel?

did you compiled your custom. generic is huge and this may cause problems.