> 
> 
> >> This seems like a serious enough problem that it should get some serious
> >> attention (are I and the couple of people who empathized the only people
> >> who ever swap?);
> 
> >Nope.  I was running crack the other day on Sparc and noticed the problem
> >for the first time (I guess I've been pretty lucky?).  The machine didn't
> >respond to anything for some 30 seconds.. then the operation returned to
> >normal.  Annoying.
> 
> And, nope again.  I notice it whenever I have GNUS (emacs-based news
> reader) running and spawn an ispell (6MB resident).  Swap, wait, swap,
> wait, swap, wait...  But, it's short-lived, and little else on my
> system causes that degree of swapping.  So, it's mearly a mild
> annoyance for me.  I can see how it could grow into a very real
> problem in a production environment, however.
> 
Indeed it is... We've seen serious problems with any significant
sendmail traffic, and web servers under netbsd have had similar
problems. Maybe we shouldn't really be trying to use NetBSD in a
production environment, but well.. the alternatives aren't too pleasant.
:) 
>