------- Original Message -------- Subject: Re: loopback Date: Sat, 12 Aug 2000 11:51:34 -0700 From: John Sage <jsage@finchhaven.com> Organization: FinchHaven To: Elisabete Reis <elreis@dei.uc.pt> References: <4.2.0.58.20000812161506.0202d890@eden.dei.uc.pt> Hi Elisabete.. I hope others more knowledgeable than I will jump in here, but here are some thoughts, and probably no answers ;-( First of all, what are you trying to do, here? Elisabete Reis wrote: > > Hi, > > I have one host with two network interfaces, eth0 (192.168.100.1) and eth1 > (192.168.170.1). OK: from this I would say you have one box that is connected to two subnets: the interface eth0 reaches the x.x.100.x subnet, and the interface eth1 reaches the x.x.170.x subnet. Presumably these two interfaces have been properly configured with ifconfig and are up. > I want to use eth0 as sender and eth1 as receiver after the date cross my > private network. I'm not sure what you mean here. If eth0 is to be "sender" only, and eth1 is to be "receiver" only, that would have to mean that the two subnets connect to each other off in the background somewhere, and in fact you have two subnets that are a closed loop. Right? Or am I missing something entirely, here? And from this it sounds as though, for some reason, you want the data to go in one direction only, creating a one-way street, as it were.. > > I have create the folowing route > > route add -host 192.168.170.1 gw 192.168.100.254 dev eth0 > > to force the data out to 192.168.100.254. Well, I think this tells the routing table to add a new route to the host 192.168.170.1 (which is *device* eth1, and is on subnet x.x.170.x) and to reach that host through the gateway with the IP 192.168.100.254 via device eth0... ...but, above, you've got the interface eth1 *itself* assigned the IP 192.168.170.1, and then in your route command you're adding that IP to the routing table as a *host* What is the IP of the specific box that these two interfaces are on, the box we're talking about, here? And where *is* 192.168.100.254 in all of this? Is this another host out on the x.x.100.x subnet? Has your routing table been set up so that the box we're talking about here knows that the host 192.168.100.254 exists? > > But when i issue a traceroute to 192.168.170.1 the date go immediately to > eth1 and no to 192.168.100.254. Yes. When you traceroute to 192.168.170.1, you're tracerouting to the interface eth1 as the destination, so packets are going to stop there. Can you ping or traceroute to any hosts out on the x.x.170.x subnet? > > Can i resolve my problem by desactivating the loopback? > If i can, how can i do it? The loopback is an interface on the specific box you're working on, and I think it's pretty much required. Each individual box has it's own loopback interface, and a lot of things depend on it being present and working... I just went back up through what I had written above, and it just occured to me: are you trying to create a hardware loopback through two NIC's, and there are *no* subnets out there? I'm still trying to envision what you're trying to do, here... > > Another solutions are appreciated. Sorry that I only asked more questions.. - John -- John Sage FinchHaven, Vashon Island, WA http://www.finchhaven.com/ mailto:jsage@finchhaven.com And remember: it's spelled l-i-n-u-x but it's pronounced "Linux" - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.rutgers.edu
