dominos

Hi.

I am not sure where to post this but I feel this is the appropriate section. [Mods, please move the thread to the right section if you feel otherwise.]

I seem to have a problem running a second instance of OpenVPN using a secondary IP address on my virtual machine (VPS) to successfully route my internet traffic through it. The first instance works well, with the OpenVPN server managed to create a tun1 adapter and route clients' traffic to the primary ethernet adapter and IP address on the machine, which is eth0 (aaa.aaa.aaa.aaa). All is well with the first one.

However, the same could not be done with the secondary IP/ethernet adapter, which its adapter name is eth0:0 (bbb.bbb.bbb.bbb) and a tun0 adapter. OpenVPN managed to bind to the said IP and I was able to connect to it and ping the internal VPN IP network (10.11.0.0/24), including the gateway (10.11.0.1) and my assigned IP (10.11.0.6), but for some wierd reasons, I could not browse the internet through it, as if the iptables was not set up correctly. It could be, given that when I try to set up the rules for the second server, I got a warning message about the ethernet adapter having a 'wierd character' in it. Although it shoots out a warning, the rule does seem to be applied when I invoke a list command ("iptables -t nat -L").

I even tried replacing the "-o eth0:0" switch to use IP address instead ("-d [IP]") but to no avail as well.

Any ideas on why it does work on the first OpenVPN server but not the second one? My machine is a Xen-based VPS, running Debian 64-bit. Replies are very much appreciated.

Here are my configuration files for your reference:

---
first-openvpn-server.conf
---
---

---
first-openvpn-server-client-config.ovpn
---
---

---
second-openvpn-server.conf
---
---

---
second-openvpn-server-client-config.ovpn
---
---

IPTables rules:

* net.ipv4.ip_forward=1 is uncommented in /etc/sysctl.conf