To: None <macbsd-general@NetBSD.ORG>
From: noud de brouwer <email@example.com>
Date: 08/23/1995 03:18:14
If i dial into BSD using ppp with proxyarp it creates 2 routes like:
pppclntone.KnoT. dns.KnoT.nl UH 0 36 ppp0
pppclntone.KnoT. 0:0:94:60:71:40 UHLS2 0 0 ae0
Is it ok to get 2 routes like that? Everything works -hmm(flow control)- ok.
But when the peer leaves (with an open enquiary for some packages??),
sometimes a route is left behind like:
pppclntone.KnoT. link#1 UHL 0 2 ae0
The second time someone come's in and starts the link pppd won't accept
pppd: ioctl(SIOCAIFADDR): Address already exists
I'm getting blue fingers in starting macro's like route delete pppclntxxx.
First i thought to do a route change or delete but only superusers can do
that and i can't su someone in a script. Moreover i don't think that's the
way to do it.
So how come there's a route left, shouldend pppd clean that for me? Did i
And, what can i do to solve it or work around it?
the "normal" settings in /etc/netstart won't prevent the route comming back.
Neither does the following settings prevent it, they make it worse.
on the "gateway" (ppp dialout on dedicated line):
on the ppp-server machine:
rarpd_flags=NO on both
and also with:
rarpd_flags="-a" on the ppp-servers
and even this showed up:
Destination Gateway Flags Refs Use Interface
193.78.85 (173) beef dead be UC 0 0 ae0
btw: i'm glad to see a "ppp-movement" on the list!