> 
> > > I think this problem is real, considering the number of people who have
> > > reported it.
> > 
> > It certainly is real. It's an old problem in the VM code.
> 
> And, unless i'm mistaken, it was one of the first problems that the
> FreeBSD folks fixed in their VM code.  (Granted the fix was a hack...
> but which is better: "hacked" or "broken"?)
> 
Actually our first "fixes" helped the problem a bit and were work-arounds.  We
now have an algorithmically complete fix.  NetBSD could probably
make the problem go away for the majority of applications (except for the
heaviest ISP-type loads), with a few simple mods.  The current FreeBSD fix
required a rewrite of the swap_pager (which needed it badly anyway) and
algorithmically finishing the collapse code (part of the problem is that
it wasn't complete.) 

If anyone wants to try their own solution then I am certainly willing to help
with some been there/done that advice.

John
dyson@freebsd.org