Linux - NetworkingThis forum is for any issue related to networks or networking.
Routing, network cards, OSI, etc. Anything is fair game.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
I have a Redhat linux box set up as a OpenVPN server. It has 2 NICs and the tunnel interface. One NIC is for the internal subnet and the second for the public interenet. I have an Ubuntu client that connects via OpenVPN. The connection comes up but the client cannot connect to any IP addresses on the server or on the internal subnet. I fired up wireshark. The OpenVPN server is seeing the packets from the client but it's sending ARP whois packets for the clients IP address. Only problem being that it's sending them on the internal subnet NIC rather than the tunnel interface so it's not getting any replies. The server's default route goes to a separate firewall router on the internal subnet that also accesses the public internet. Not sure if thats part of the problem but I don't see how it would be.
The ultimate goal is for the client to have complete access to the server internal subnet. Currently I have the firewall pretty much shut off on both the internal and tunnel interfaces. Below is all the configuration info I think is pertinent.
The big question is why the ARP packets are being broadcast on the internal subnet NIC when there is a route going over the tunnel interface for the IP address of the client.
Not sure what the hell I've got screwed up. Any hints would be greatly appreciated.
I've always wanted to play with OpenVPN, but haven't taken the time. Forgive me if I'm way off base, but in looking at this, I'd expect both ends of the tunnel to have matching IP addresses, and it looks like yours don't:
- Is there a reverse route on the internal machines to reach the machine connected by VPN? Looking at your setup 10.91.92.X will not be reachable by 10.91.91.X since they are on different subnets. Without a route on 10.91.91.X to reach 10.91.92.X your setup will never work.
- Have you enabled routing on this VPN Server? cat /proc/sys/net/ipv4/ip_forward should return 1.
- Is there a reverse route on the internal machines to reach the machine connected by VPN? Looking at your setup 10.91.92.X will not be reachable by 10.91.91.X since they are on different subnets. Without a route on 10.91.91.X to reach 10.91.92.X your setup will never work.
- Have you enabled routing on this VPN Server? cat /proc/sys/net/ipv4/ip_forward should return 1.
Thanks for the input. There are routes set up on the internal machines but the client can't even talk to the server which should be part of both subnets and do the routing between them. And yes ip_forward is set to one. It was easy to miss but I included the output from sysctl -a and also it's set to one in /etc/sysctrl.conf. Thanks again and any other ideas are appreciated.
I've always wanted to play with OpenVPN, but haven't taken the time. Forgive me if I'm way off base, but in looking at this, I'd expect both ends of the tunnel to have matching IP addresses, and it looks like yours don't:
Of course, having an all-ones mask looks odd, too, but I assume that's part of the 'P-t-P' magic.
ch
Hmmm...not off base at all. Don't know how I missed that. Not sure if it's the cause but it certainly looks like something to check. The interesting part is that all those IPs are generated dynamically by OpenVPN. Gonna have to try to figure out how it comes up with them. Anyway, thanks for the input.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.