growlf
Hello,
Long/Short story. My elderly parents live in the middle of nowhere with ZERO wired Internet options. Amazingly, they have decent ATT cellular data access. Unfortunately, ATT cellular connections do all sorts of weird proxying and allow no incoming connections. To get around this, I've had them using OpenVPN on a Tomato router connected to a StrongVPN account (no encryption, using TCP w/ compression - that's it). Works fine. I've got miniupnpd properly working with the connection, port forwarding, etc - in essence the OpenVPN connection makes everything appear like a nice normal network connection.
Earlier this month I decided that I may as well host my open OpenVPN server on a VPS. I'm not convinced that the StrongVPN servers aren't overloaded, and I'd like to be able to configure the system myself.
So, the behavior I "see" with StrongVPN is that I connect to the IP address that I consider "my" static IP address. Once connected, all traffic to that IP is sent to my router and I can handle masquerading, port forwarding, etc on the router (presumably MINUS the ports used for the connection itself).
Looking at setting up a server myself, it appears that if I have multiple IPs on my host, I can set a static public IP for the connection that ISN'T the IP of the OpenVPN server. This seems like it would be the cleanest approach, but would require multiple IPs (something I don't have with the VPS provider I'm testing). So, my question is - to create something akin to the behavior I'm seeing with my StrongVPN connection, can someone tell me a few of the key concepts I should be focusing on in the OpenVPN setup?
I know this is utter newbie-ism, and I will happily RTFM - but I'm hoping to avoid going down any roads that will just lead to frustration. If I can do this with a single public IP - SUPER. If not, I'll make sure I'm setting this up somewhere that can accommodate.
Thanks very much for taking the time to read this.
Best wishes,
john
growlf
I'm still reading, but so far not seeing an obvious solution...
I'm good with routing all client traffic thru the VPN. No problem. I setup the masquerading rule on the server and I'm good to go.
What I want to be able to do, however, is setup the server so that I can manage incoming connections to its IP and forward them to machines behind my router's VPN client. Perhaps I would forward all traffic to the private IP of my router's side of the VPN connection and then have it forward to the inside machines. Double NAT'ed?
It took a bunch of screwing around, but the way I've got things working with StrongVPN now, client machines (behind my tomato router) can make forwarding requests with UPNP and PMP and they "just work". If each client connected independently to the VPN server, I could presumably just do forwarding on the server - but that doesn't get me to the level of "just works" that I have now.
Looking at my strongVPN client config, the server I connect to has the same IP that I use as my "external IP" once the connection is established, so it would seem likely to me that I'm just dealing with a double-natted scenario with them as well - but I'd really like to know what their forwarding rules look like on the server side.
Anyway, thanks for listening to the ramblings of a person with zero practical OpenVPN experience!
--- John
tristan_c
Did you ever get this sorted? I'm trying to achieve exactly the same...
I have a VPS with a static IP and single NIC. My plan was to forward all traffic apart from ports 22, 80, 53 etc to the tunnel (and hence the router), then forward a high random port to port 22 on the router so I can get 'direct' access.
No idea if this is possible.
T
