Why ue0 do ARP on non local address when using static route?

Lowell Gilbert freebsd-questions-local at be-well.ilk.org
Wed Feb 13 22:38:48 UTC 2013


Jin Guojun <jguojun at sbcglobal.net> writes:

> This is 8.3-Release on a HP EliteBook 8460p (4-core i5) with an on board Intel 
> (em0) interface.
> When attached a Trendent TU2-ET100 USB Ether dongle for a second interface, it 
> has no problem to talk to the local network (10.234.37.0/24), but it has problem 
> to talk to a remote network or host (10.227.148.0/24) via eu0 interface.
> When a remote host ping this host or this host ping that remote host, ARP 
> request is always showing up.
> A static route is set and remote host is no part of the local sub net, why ARP 
> is going on?
>
> Is any sysctl parameter can fix this problem?

Until we know what the problem is, we can't guess at the solution.
You didn't show the commands you used, for setting up the routes *or*
for the diagnostic information you provided.

> -Jin
>
>
> Internet:
> Destination        Gateway            Flags    Refs      Use  Netif Expire
> default            10.234.16.1        UGS         0  2841993    em0
> 10.227.148.52      10.234.37.80       UHS         0       26    ue0

There's no "G" in the flags there. 
Sign of a problem.

> 10.234.16.0/22     link#1             U           0        0    em0
> 10.234.17.41       link#1             UHS         0        0    lo0
> 10.234.37.0/24     link#8             U           0        3    ue0
> 10.234.37.80       link#8             UHS         0        0    lo0

That's the gateway address. What's it doing with a host route on the
loopback interface? That will override the 10.234.37.0/24 interface
route, and turn 10.234.37.80 into a "black hole."

 - Lowell


More information about the freebsd-questions mailing list