OpenVPN client in Linux Containers
This article describes how to setup a Linux Container to run OpenVPN in client mode with a "kill switch" for secure/private internet use. Doing so offers a distinct advantage over using full-blown virtualization like VirtualBox or QEMU in that the resource overhead is minimal by comparison and able to run on low powered devices.
Container setup
Basic setup and understanding of Linux Containers is required. This article assumes that readers have a base LXC setup operational. Newcomers to these are directed to the aforementioned article.
LXC configuration
The container's configuration should be modified to include several key lines in order to both run OpenVPN and have internet programs (browsers, email clients, torrent clients, etc.) interact with the host system from within the LXC and from behind the VPN.
For the example, the lxc is named "playtime" and a full configuration is shown:
/var/lib/lxc/playtime/config
... ## for openvpn lxc.mount.entry = /dev/net dev/net none bind,create=dir lxc.cgroup2.devices.allow = c 10:200 rwm
Needed packages within the container
In addition to the base system, openvpn is required and available from the official repositories. A properly configured firewall to run within the container is highly recommended. The role of the firewall within the container is two fold:
- Provide a functional "kill switch" to maintain privacy should the connection to the VPN fail.
- Keep nasty stuff out.
This guide uses ufw which is very easy to configure, but other examples can certainly be used.
Package setup
OpenVPN
Configuration of OpenVPN is beyond the scope of this article. Readers are encouraged to read the OpenVPN article to properly setup the software for a given VPN provider. Note that many private VPN providers include links to directly download a properly configured openvpn.opvn profile unique to their particular service. For the purposes of this guide, /etc/openvpn/client/myprofile.conf will refer to that config.
Verify openvpn functionality within the container; start openvpn via [email protected] and once satisfied enable it to run at boot.
LimitNPROC...
Avoiding DNS leaks
Users are highly encouraged to setup openvpn with to manage /etc/resolv.conf via the instructions in OpenVPN#DNS. Failure to do so may lead to DNS leaks if the host's DNS server is specified in said file.
ufw
Configuration of ufw is described in OpenVPN#Firewall configuration. Once modified as described, additional setup is simply to define the protocols to run (ssh, torrent, etc.) and also define the IP address or address range of the VPN provider. Again, these are available from the private providers.
First setup the default deny policy and then allow whatever services are to pass though:
ufw default deny outgoing ufw default deny incoming ufw allow ssh ufw allow 1194 ufw allow out 1194 ufw allow out on tun0 from any to any ufw allow in on tun0 from any to any
Now add the VPN server IP addresses or ranges. Note two entries for each IP address is required. In the example below, only a single fictitious IP address is shown for illustrative purposes.
ufw allow in from 50.120.10.200 to any ufw allow out from 50.121.10.200 to any
Finally, add the internal LAN IP range to allow access regardless of VPN connectivity:
ufw allow from 192.168.0.0/24
Start ufw and enable ufw.service to start at boot.
# ufw enable
pgl
Additional protection can be had via using pgl within the container. See: PeerGuardian Linux#Running pgl from within a container.
Test the service
From within the running container, (connected via ssh or via lxc-attach -n playtime) test the setup by exporting a browser to the host's machine X server:
$ DISPLAY=:0 firefox
The result should be a firefox window in the host's X server with the title, "Mozilla Firefox (playtime)." A number of websites can be used to verify IP address and status of DNS entries. Once such site is ipleak dot net.
At this point, the IP and DNS entries corresponding to /etc/openvpn/client/myprofile.conf should be displayed.