weird cause of timeouts for apache

[David Vincelli <leao@sympatico.ca>]

	After I upgraded to OpenBSD 2.8 (i386), my web server became
nearly unresponsive from requests made on the internet.  On my local
network, however, everything was as reliable as ever.

	Recently I realized that I was using the wrong MTU size for
my DSL provider.  After adjusting the MTU size to the proper value
(1454), I noticed some services that used to timeout now work 
properly.  I figured this is what was plaguing apache.  To my
surprise, it didn't make a difference.   Interestingly enough,
after I run 'route change default -mtu 1454' as root, my web server
becomes accessible from the internet.  

	The question(s) that arise from this little glitch, are
is this a bug in apache?  Or maybe in pppd?  I use pppd + pppoe to 
establish my connection with my provider.  I tried to write a little
suid c program to run from /etc/ppp/ppp.linkup but route complains
that it's not being run by root.  I have no experience writing suid
programs so that might be my fault.  All my c program contains is 
a single execl( ..., ... ); statement, if I change /sbin/route to
/usr/bin/id, I can clearly see the euid is 0.  

	Thanks,

--
David Vincelli <leao@sympatico.ca>