Date: Fri, 24 Jan 1997 13:43:57 +0200 (EET) From: Andrew Stesin <stesin@gu.net> To: John Polstra <jdp@polstra.com> Cc: hackers@freebsd.org Subject: Re: Fault-tolerant network with 2 ethernets Message-ID: <Pine.BSI.3.95.970124130257.5513E-100000@creator.gu.kiev.ua> In-Reply-To: <199701232316.PAA21713@austin.polstra.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 23 Jan 1997, John Polstra wrote: > The point about 100BaseT failing differently from coax-based ethernet > (the whole net doesn't go down if there's a cable problem) is well > taken. You still have a hub, though -- hub's failure is similar to a coax' failure, everything on the segment dies. > > Non-simple solution. Three addresses for every machine. one on > > each ethernet segment, and one off the loopback address. That's a usual way of doing things on routers -- router-ID is an alias tied to the loopback. > ifconfig lo0 inet 192.168.3.21 alias netmask 0xffffffff > ^^ > || > host-specific, of course Yes. But make sure that a set of IP addresses used for loopback-aliased router-IDs is _not_ covered by any of the subnets (address ranges). > At the application layer, the 192.168.3.x addresses are the ones > that would be used. Yes, probably. But just to be sure, I'd prefer to use gethostbyname() and try all the addresses for the given hostname in turn -- or even simultaneously. > The applications would never use the real addresses of the ethernets. Yes. This may be achieved just with a proper DNS: Direct zone: host1 IN A 192.168.3.1 ; do we really need the 2 following? ; IN A 192.168.1.1 ; IN A 192.168.2.1 Reverse zone: 192.168.3.1 IN PTR host1.dom.ain. 192.168.1.1 IN PTR host1.dom.ain. 192.168.2.1 IN PTR host1.dom.ain. > Finally, each host would advertise a route to its 192.168.3.x > address on both ethernet interfaces. In effect, each host acts as a > gateway to itself. Yes. OSPF will do this for you almost automagically -- you just use a "stubhost" to describe this alias (router-ID) in "ospf" section of gated.conf. > I read up on OSPF and gated this morning, and it looks like they're > perfect for solving this problem. The documentation even seems to > suggest the trick of adding a second address to the loopback > interface, if I understand it right. Yes. This is a usual approach, as I noticed already. -- Best, Andrew Stesin nic-hdl: ST73-RIPE
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSI.3.95.970124130257.5513E-100000>