server-server

NETunnel Server-Server example

In this example we'll create 2 instances of NETunnel servers, server1 and server2, each with its own configuration file.

server2 settings will include a single peer - server1 with a single static tunnel to this peer.

There are several advantages in using server-server model instead of the client-server model. First of all, A single server instance can create more than 1 tunnel to more then 1 remote, and secondly, these tunnels are designed to recreate themselves if there was a disconnection.

Another useful advantage is that peers and static tunnels can be dynamically created using the NETunnelClient.

Usage

Run the first server:

python -m netunnel.server -p 4040 -c server1.conf

Run the second server:

python -m netunnel.server -p 4041 -c server2.conf

The server2.conf configuration file include a peer called server1 with a static tunnel to it on port 20000. Let's make sure it works:

ssh -p 20000 localhost

Notice that if you stop the instance of server1 from running, the second instance will try to reconnect it.

Now, make sure we have both server1 and server2 running and let's use server1's API to add server2 as a peer and a static tunnel to it:

python client.py

Now server2 is registered as a new peer and you will see that server1.conf was updated. The local port that server1 chooses for the tunnel is an available port number from 20000 or more, so we'll assume it chose port 20001 and now we can run:

ssh -p 20001 localhost

You can also see that server1.conf was updated so it will recreate the tunnel on startup.