virtualbox.org

To get a better understanding of the issue here's the scenario i'm trying to implement:

Physical Host: Windows XP

VMachine1:
Name: Berlin-HQ-W2k8
OS: Windows Server 2008
Network: Internal Network: 'Berlin-Server-LAN'
IP: 192.168.3.2/29
GW: 192.168.3.1
No DNS

VMachine2:
Name: Router1-Berlin
OS: Linux (Kernel2.4/floppyfw)
Network:
-Internal Network eth0: 'Berlin-Server-LAN'
-Internal Network eth1: 'Berlin-Munich-WAN'
IP eth0: 192.168.3.1/29
IP eth1: 10.0.01/30

VMachine3:
Name: Router2-Munich
OS: Linux (Kernel2.4/floppyfw)
Network:
-Internal Network eth0: 'Munich-Server-LAN'
-Internal Network eth1: 'Berlin-Munich-WAN'
IP eth0: 192.168.4.1/29
IP eth1: 10.0.0.2/30

VMachine4:
Name: Munich-Branch-Core
OS: Windows Server Core
Network: Internal Network: 'Munich-Server-LAN'
IP: 192.168.4.2/29
GW: 192.168.4.1
No DNS

I created static routes between Router 1 and Router 2 on the linux machines so that the 192.168.4.x / 192.168.3.x are reachable. Pinging the non-directly-connected LAN interfaces works on the software routers. So i guess the static routes should be in place. The problem is that the Windows machines can't reach each other. They can only ping the router local to their site (LAN+WAN interface) but every other ping reaching the opposite site fails. Firewalls are off. Wireshark returns a header checksum 0x0000 error code when sniffing the NIC on the Win2k8 Server. I tried differnet Linux distros, but it wouldn't solve the problem either. I hope there is a solution to this problem...

Internal network 1 can not communicate with internal network 2, this is normal behaviour.

vbox4me2 wrote:Internal network 1 can not communicate with internal network 2, this is normal behaviour.

Actually this is not normal behavior, since he has a router connected on both networks.
How are you trying the connection between the machines? Simple ping?
Are you sure the windows firewall on the Server machines are turned off? Did you try a traceroute? Where does the packet dies?

I've a similar setup as yours ( LAN <-ROUTER-> LAN) and all works fine. So this is probably a configuration issue on the guest VMs.

You say that you added static routes, can you post them here. I'm sure you made an error somewhere, like adding only routes to the routers and not the other machines. This is basic network solving, draw the whole situation, note the IP addresses at each line and check the routes. You have to make sure that there is a route TO the machine, as well as BACK from it. The latter is a common mistake.

Route on Router1:

ip route 192.168.4.0 netmask 255.255.255.248 10.0.0.2 dev eth1

Router 2:

ip route 192.168.3.0 netmask 255.255.255.248 10.0.0.1 dev eth1

So when it comes to configuring Cisco routers these are the only steps necessary. And this works on real hardware
And a default route doesn't make sense, does it?

Traceroute instantly dies with a timeout when trying to reach the opposite site...

Even on Cisco Routers, this isn't enough. Like I said, you need a route BACK to the source. The traffic knows it's way to the other machine, but that doesn't know it's way back.

My understanding of static routes is that you only tell the router those networks which are not directly-connected. In the case of Router1: Network 192.168.4.0 is not known because it's connected to Router2. So you have to tell R1 the way to reach that network over a directly-connected one on that router (10.0.0.0 reaches 192.168.4.0). The same thing on Router2 (10.0.0.0 reaches 192.168.3.0). So which additonal route is missing??