My company operates a supermarket chain. We have recently brought all of our gift card and check processing in house. We have one server at our corporate office that performs both of these tasks.
We really want to get some redundancy in place in case there is a problem with the connection at our office or if there is a problem with the server. The best solution was to have a redundant server at one of our stores. We got the server, got it installed, set up the database replication, but now we've come to the hard part of the setup.
The networking part of this project is a little tricky. The IP range of the corporate office and the IP range of the store with the redundant server are different. Unfortunately our point of sale system only allows us to set up on IP address per tender type. If we could enter a fail over IP in the POS then this would be easy, but we can't. Since we can only have one IP address per tender we need to have the main and redundant servers set up as the same IP address.
We originally thought that we would add a second router to the network at the redundant site that would have the same internal range as the range at the office and then add an additional route to all of the routers at the other stores to the redundant server with a lower weight than the route to the office. I would then set the 2nd NIC in the redundant server to the same IP as the server at the office and route traffic accordingly if the connection at the office dropped or if the server at the office was down. There was a problem with this design. Everything is fine and dandy until I hook up the 2nd NIC to the 2nd router at the redundant site. After doing that the two servers are unable to communicate. Logically this makes sense to me. A packet is sent from the main server (192.168.100.3) to the redundant server (192.168.112.3) and when the redundant server tries to send back the ACK to the main server at 192.168.100.3 it just gets routed to the 2nd NIC of the redundant server. So in the case of a failure of the server or the connection at the office, this design works, but I'm unable to replicate my data to the redundant server.
I've tried setting up a static route on the redunant server telling it to send all traffic destined for the office out of the interface with the 192.168.112.3 IP and bypass the 2nd router entirely. It didn't work.
I've sat and thought about how to get this to work as an automatic fail over and I'm stumped. The only way that I have figured out to get this to work would be to call someone at the store in the case of a failure and have them plug in the cable to the 2nd NIC. This is obviously not an automated process and that's what we're trying to achieve.
Anyone have any good ideas about how we can get this to work?