At the moment I have 4 squid caches connected to a single Foundry Networks ServerironXL via 100Mbit full duplex links. The ServerIron is connected into our backbone network via Gigabit link.
Each of the squid caches has 2 100Mbit ethernet ports on the motherboard
As the above confgig has a single point of failure (the ServerIron) what I'd like to do is implement a dual homed setup using two ServerIrons and have each cache connected to each ServerIron. This, according to the Foundry docn, will allow me to create an active-active setup where traffic from our net will be able to access all our squid caches via either ServerIron.
All the caches and ServerIrons will be on the same class C subnet.
The problem I've got is that while inbound traffic will be load balanced over the 8 fast ethernet ports on the cache boxes, I don't know how to persuade squid to load balance outgoing traffic over the two interfaces on the server. Normally I'd just have a default gateway setup on each cache but that's not going to work in this case as I'm trying to use two interfaces.
So ... if i've got 2 interfaces on a squid cache on the same net, how do I persuade squid to load balance outgoing traffic over both interfaces.
TIA Alex
