[lug] forcing certain services to use eth1 instead of eth0

Michael Belanger mrb at ciclops.org
Tue Jan 3 16:58:14 MST 2006


On Tuesday 03 January 2006 16:55, Nate Duehr wrote:
> Hang on... seems like you're trying to get too fancy here.
>
> I'm assuming that you've got two different IP addresses on eth0 and eth1?
>
> If so, then on the "inbound from clients" connection -- they're
> connecting to the IP on eth1 -- nothing to set up there.  Your router
> and/or switch should be ARP'ing for that IP and sending the traffic to
> the physical port that answers the ARP request.  No-brainer.
>
> On the "outbound toward client" side of the connection... your daemon
> should be answering using the IP address (socket, really) it received
> the connection on?  If so... it'd all be going out eth1 also.  If the
> application isn't doing that, something's not right.
>
> So... I'm not sure why you'd need any firewall rules or anything other
> than the correct IP's in everything, including the daemon configurations
> for whatever is listening on your machine.  Traffic to eth0's IP would
> use eth0... traffic to eth1's IP would use eth1...???
>
> Nate

I am all for making things simpler.  The eth1 is a seperate IP but on the same 
network.  However, it is a new connection.  Perhaps I just need to clear the 
routing table?

-Michael
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.lug.boulder.co.us/pipermail/lug/attachments/20060103/bec2389c/attachment.pgp>


More information about the LUG mailing list