Subject: Mad route and can't allocate llinfo
To: None <tech-net@netbsd.org>
From: Stephen Borrill <netbsd@precedence.co.uk>
List: tech-net
Date: 02/25/2004 14:47:38
I've got a server running 1.6.2 with two network interfaces. fxp0 is on 
a 10.0.0.0/16 range and fxp1 is on an external IP address range. The
machine is not running NAT or acting as a gateway. The machine is
frequently reporting the following error:

arpresolve: can't allocate llinfo on fxp0 for 255.255.0.0

As this usually indicates a gateway that isn't on a local network, I
looked at netstat -rn:

Internet:
Destination      Gateway           Flags     Refs     Use    Mtu
Interface
0&0xa006402      255.255.0.0       UGS         0    14513      -  fxp0 =>
default          a.b.c.d           UGS        10   558927      -  fxp0
10/16            link#2            UC         17        0      -  fxp1
10.0.0.1         00:01:03:d2:cf:78 UHLc        0       39      -  fxp1
10.0.100.2       00:03:47:4b:8f:ff UHLc        1        2      -  fxp1

The top entry looks pretty weird to me. 0xa006402 is hex for 10.0.100.2
(for some value of byte order). Where might this entry have come
from? How can I delete it? What does the => mean after the interface name?

There are other (mainly Linux) machines that I do not administer similarly
connected between the two networks. Might these machines be doing
something to cause this?

-- 
Stephen