Subject: Re: routing problems across disjoint segments
To: Paul Newhouse <newhouse@rockhead.com>
From: Manuel Bouyer <bouyer@antioche.lip6.fr>
List: netbsd-help
Date: 03/31/1999 16:17:45
On Mar 31, Paul Newhouse wrote

[...]

> newhouse netstat -nr looks like:
>    Internet:
>    Destination        Gateway            Flags     Refs     Use    Mtu  Interface
>    default            205.219.89.41      UGS         5     6482      -  de1
>    24.1.4.192/27      link#3             UC          0        0      -  ne0
>    24.1.4.193         24.1.4.200         UGHS        0      221      -  ne0
>    24.1.4.200         00:c0:26:38:80:71  UHL         1        0      -  lo0
>    127.0.0.1          127.0.0.1          UH          1      123      -  lo0
>    205.219.89.40/29   link#2             UC          0        0      -  de1
>    205.219.89.41      00:20:6f:02:fa:bb  UHL         3        0      -  de1
>    205.219.89.42      00:40:05:a0:4e:b2  UHL         2     3007      -  de1
>    205.219.89.45      00:40:05:a0:41:2a  UHL         0       20      -  lo0 =>
>    205.219.89.45/32   link#1             UC          0        0      -  de0
>    205.219.89.46      00:40:05:42:35:d0  UHL         0        0      -  lo0
>    206.100.6.73       205.219.89.41      UGHS        0        0      -  de1
>    206.100.6.74       205.219.89.41      UGHS        0        0      -  de1
> 

This may be the cause of problem #1:
your routes to 205.219.89.4{1,2} go out by de1, where they should use de0.

This problem is not easy to solve, because it involves arp too.
What I'd try is: ifconfig de1 with a /32 netmask, so that all your arp entries
would add a route via de0 (note that with your current config, your
205.219.89.40/29 network is marked reacheable on de1, which is wrong),
then add a specific route for 205.219.89.41 (possibly using the -iface
modifier).

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--