I had been running NetBSD1.4.1 for several weeks, using a 4Gig Seagate
Barracuda Drive on the same system.  As I recall, I got a few crashes of
sorts when using the package source procedure to download, compile, and
install various packages.  Recently, I transferred the system to the current
drive, which is a 4Gig Medalist Drive.  The drive was already installed with
1.4.1, and I simply copied the etc, usr, and home directories.  Everything
works well, except for this problem.  I am not sure whether it is a bug or
simply user error in setting up the system.  Hence, my query regarding the
experience of others.  If this is a real bug, and not a system set up
problem, I would expect the problem to be common.

I have submitted a bug report.  However, I have never used the reporting
procedure before, so I am not sure it got through.

Jim Nelson

-----Original Message-----
From:	netbsd-help-owner@netbsd.org [mailto:netbsd-help-owner@netbsd.org] On
Behalf Of Todd Whitesel
Sent:	Monday, November 15, 1999 12:23 AM
To:	jimtnelson@earthlink.net
Cc:	netbsd-help@netbsd.org
Subject:	Re: pagedaemon panic

> I get this error when running printing several graphics with gnuplot,
using
> the apsfilter package, which involve generatign postscript files, and
> ghostscript output to the printer.
>
> I also got this crash when running a batch job and simultaneously
compiling
> and linking a fortran code.
...
> I am running NetBSD-1.4.1/i386 on a 133 MHz Pentium with 16 Meg RAM and a
> Seagate 4Gig drive, with an Adaptec 2940UW.

Hmm, I've got a 333 mhz PPro (overdrive PII, yes, they make one for socket
8)
with a 2940UW on the motherboard, an IBM 2gig and seagate 4gig, and I get
occasional total freezes from 1.4.1/i386 -- always when using lots of
memory,
apparently soon after things start to use swap (according to 'top').

For me it hits when running lots of CVS checkouts from other machines to the
repository on this machine (rsync'd NetBSD anoncvs) and also when viewing
very
large TIFF scans with xv. And the last time I built a new 1.4.1 kernel in
text
mode, I got a spurious segv in cc1.

Smells related to me. I suspect a subtle bug in UVM. I'm running a lightly
trimmed kernel on this guy so if there are good debugging ifdef's for this
sort of thing, I could enable them.

Todd Whitesel
toddpw @ best.com