Lixonet Enterprise Edition (LIXONET 3)

This is the semi-official, testing Lixonet EE repository.

Custom configurations should be supplied in a lixonet.conf file in the name=value format, such as your subnet. Additionally, a tinc keypair is also be required.

Setup, Installation, and Updating

Prerequisites/Setup:

1. eth0 is the LAN or WAN interface supporting external routing, DNS, etc., and is the bridged interface to a router that will statically route the entire desired network subnet (i.e. x.x.0.0/21) through it. You may also have an eth1 (and so on, so forth) that you statically configure for your entire, wide subnet (i.e. x.x.0.0/21) if you want to use two physical adapters. If you do have more than one interface, make sure that internal_gateway is set to the router IP address that can route your personal Lixonet subnets, and not your WAN gateway (unless they're the same).
2. tun and tap are in /etc/modules to load at boot (https://www.cyberciti.biz/faq/linux-how-to-load-a-kernel-module-automatically-at-boot-time/) and the system has been rebooted afterwards.
3. IPv4 forwarding is on: sysctl -w net.ipv4.ip_forward=1
4. Clone this repository: git clone <url>
5. mkdir /etc/lixonet/(mesh)/ where (mesh) is something like teamlixo if you're planning to use teamlixo.mesh.
6. Fill out /etc/lixonet/(mesh)/lixonet.conf (see: Configuration options)
7. Supply /etc/lixonet/(mesh)/tinc.key (see: Generating keypairs for tinc)
8. Supply /etc/lixonet/id_rsa and /etc/lixonet/(mesh)/id_rsa for each mesh you'll be connected to (see: Generating SSH keys). Remember each id_rsa is unique!
9. Supply /etc/lixonet/known_hosts (see: Setting known_hosts)

Generating keypairs for tinc (tinc.key):

Source: https://www.tinc-vpn.org/documentation/Generating-keypairs.html

Run: tincd -n lixonet -K

Generating SSH keys (id_rsa):

1. I recommend PuTTYgen (https://www.puttygen.com/) on Windows, ssh-genkey on Linux.
2. Generate at least two keys. One will be for lixonet-ee, one will be for whichever repositories you're a part of (i.e. teamlixo.mesh).
3. Share the public keys (starting with something like ssh-rsa) with the repository owner(s) listed below so they can add a "Deploy key" to the repository for you.

Setting known_hosts

For GitHub, at this time the contents would be:

|1|hW/UPBCtfR0M/2GejxakBvoqGNY=|DXk4SRaJXzawNdHyBe9mrxvWYvk= ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ==

Don't take my word for it, see: https://docs.github.com/en/github/authenticating-to-github/keeping-your-account-and-data-secure/githubs-ssh-key-fingerprints

Optional: Generate a WireGuard key

You only need to do this if you are setting wg_enabled=1, for example when you want to service VPN clients from a "master" node. See Wireguard Options for more information on how to do that.

See: https://www.wireguard.com/quickstart/

apk add -U wireguard-tools
wg genkey | tee privatekey | wg pubkey > publickey

Install/Update:

1. Read and follow the prerequisites above and make sure you are ready to install.
2. Clone or navigate to a cloned repository folder you already have.
3. Run chmod +x install.sh && ./install.sh
4. Enjoy!

Configuration options

Location

A lixonet.conf file should be under each network directory you'd like to create. Create one directory for each mesh you'll be joining, under /etc/lixonet. If you want to join the teamlixo mesh for example, do this:

1. Create /etc/lixonet/teamlixo/lixonet.conf and specify at least all required options in this file.
2. Create or save /etc/lixonet/teamlixo/tinc.key

Available networks

Official networks:

Name	Git clone URL(s)	Subnet	ASN format	Administsration
teamlixo	http://git.team.lixo/lixonet/teamlixo.mesh.git	172.31.0.0/16	4206969XXXX	Manevolent, Nurd
colorado	http://git.team.lixo/lixonet/colorado.mesh.git	172.29.0.0/16	4206970XXXX	Manevolent

Format

All options are supplied in the name=value format on individual lines, i.e. name=test_name_lixo

Required options

You MUST specify these options, or Lixonet will not work.

• git - The Git clone repository URL, used for peer configurations.
• tinc_peer_name - The global neighbor name. Tinc and Bird use this to identify a peer in the network, and to exclude your own pre-packed config automagically from Tinc and BIRD dialing out so you don't connect to yourself over and over again.
• tinc_peer_address - The router address to use. This is your Lixonet routing layer address: 172.xxx.0.xxx
• network_address - The network address to use. This is your Lixonet routing layer address: 172.xxx.0.0 (especially take note of the last two 0's: 0.0 -- it MUST end with zeros corresponding to the network size)
• bgp_asn - The BGP ASN to use. We usually follow the format 4206969XXX where XXX is the last octet of your address, zero-padded (i.e. 008 or 212)
• internal_address - The internal address to use. MUST be unique to this Lixonet EE instance; if the address of the host is 172.31.16.2, you should NOT use that IP address, and instead pick another static address.
• internal_gateway - The internal gateway to use. This is the IP address, directly upstream from the Lixonet EE router, that can be used as a next hop to reach your own subnets.

Configuration example

A copy of the live configuration (/etc/lixonet/teamlixo/lixonet.conf) used by denco_mane_lixo:

git=http://git.team.lixo/lixonet/teamlixo.mesh.git
tinc_peer_name=denco_mane_lixo
tinc_peer_address=172.31.0.8
tinc_bind_address=172.31.8.2
bgp_asn=4206969008
bgp_routes=172.31.8.0/21,172.31.254.0/24
network_address=172.31.0.0
netmask=255.255.255.0
internal_gateway=172.31.8.1

(A key exists at /etc/lixonet/teamlixo/tinc.key)

cat /etc/network/interfaces:

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet dhcp
        hostname lixonet

ifconfig eth0:

lixonet:/home/manevolent# ifconfig eth0
eth0      Link encap:Ethernet  HWaddr xx:xx:xx:xx:xx:xx
          inet addr:172.31.8.2  Bcast:0.0.0.0  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:313268 errors:0 dropped:0 overruns:0 frame:0
          TX packets:266835 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:223824994 (213.4 MiB)  TX bytes:145971829 (139.2 MiB)

docker ps -a | grep teamlixo (built by ./build.sh):

646e3b6b5456        teamlixo_tinc       "sh -c '/usr/sbin/ti…"   17 minutes ago      Up 15 minutes                                 teamlixo_tinc_1
84fda16f829b        teamlixo_bird       "/bin/sh -c 'bird -c…"   17 minutes ago      Up 9 minutes                                  teamlixo_bird_1

route

Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
default         172.31.8.1      0.0.0.0         UG    202    0        0 eth0
172.17.0.0      *               255.255.0.0     U     0      0        0 docker0
172.18.0.0      *               255.255.0.0     U     0      0        0 br-b91440553ab5
172.31.0.0      *               255.255.255.0   U     0      0        0 teamlixo
172.31.0.0      *               255.255.0.0     !     32     0        0 *
172.31.8.0      *               255.255.255.0   U     0      0        0 eth0
172.31.8.0      172.31.8.1      255.255.248.0   UG    32     0        0 eth0
172.31.16.0     172.31.0.16     255.255.248.0   UG    32     0        0 teamlixo
172.31.200.0    172.31.0.200    255.255.248.0   UG    32     0        0 teamlixo
172.31.208.0    172.31.0.208    255.255.252.0   UG    32     0        0 teamlixo
172.31.212.0    172.31.0.212    255.255.252.0   UG    32     0        0 teamlixo
172.31.254.0    172.31.8.1      255.255.255.0   UG    32     0        0 eth0

(Note the use of the internal_gateway variable to set the local 172.31.8.0 172.31.8.1 255.255.248.0 UG 32 0 0 eth0 route!)

Non-required options:

General network options

• global_prefix: global network prefix: defaults to 16.
• router_prefix: router network prefix: defaults to 24. If you change this you SHOULD change netmask too.
• netmask: router netmask: defaults to 255.255.255.0 (/24); SHOULD be the bitmask that corresponds to the router_prefix
• internal_interface: the interface to route your OWN networks to. This should be the interface where your personal Lixonet subnets are reachable at, or in other words the interface at which the Lixonet EE router can send packets destined to networks you own. Defaults to eth0.
• internal_subnet: the internal subnet that your Lixonet EE router is connected to, defaults to "internal_gateway/24".
• vip: virtual IP: a virtual IP address to assign to the Lixonet EE router directly, aside from 172.31.0.8.

Tinc options

• network_name - The name of the interface that Tinc should create when it's started. Defaults to lixonet. Set this for multiple networks.
• tinc_bind_address - The address that Tinc should bind to to listen for incoming public Internet connections (i.e. 10.0.0.1). This is not the Lixonet router IP (i.e. 172.31.0.8). Use this if you have an eth1 that you don't want to use for binding tinc to, and would prefer to bind tinc exclusivley to eth0 instead. Defaults to internal_address.
• tinc_bind_port - The Tinc server port to listen on. Defaults to 9994.
• tinc_connect_to - A comma-separated list of well-known/pre-defined hosts to connect to (i.e. denco_mane_lixo). If not supplied, this is automatically set to all core routers that aren't yourself (name).

BGP options

For options we expose here, for information see: https://bird.network.cz/?get_doc&v=20&f=bird-6.html#ss6.3

• bgp_path_metric: Enable comparison of path lengths when deciding which BGP route is the best one (0 or 1): defaults to 1
• bgp_aigp: BGP AIGP state (enable, disable, or originate): defaults to originate (see BIRD documentation)
• bgp_rpki_retry: If RPKI cache data cannot be obtained, the time period in seconds between a failed query the next attempt. Defaults to 90.
• bgp_rpki_refresh: How long to wait in seconds before attempting to poll RPKI cache data after the last successful poll. Defaults to 900.
• bgp_rpki_expire: How long to keep any records locally cached before they are deleted. Defaults to 172800 (2 days).
• bgp_rpki_known_hosts: The file path for the SSH key known_hosts file to use when validating remote RPKI hosts. Defaults to /etc/bird/rpki/known_hosts (provided by Lixonet; don't change this unless you need to!).
• bgp_passwd_*: The password for a specific BGP peer (i.e. bgp_passwd_denco_mane_lixo). If not specified, no password is used for the host.
• bgp_routes: A comma-delimited list of route prefixes (CIDR) to advertise to other BGP peers. If not specified, no routes are propagated.

BIND options

These settings are optional, and are already defaulted to general network common practices.

• bind_forward_enable: Enable BIND DNS forwarding when DNS queries are received from other nodes Lixonet for a domain that you control (i.e. lkwco.mane.lixo on that Lixonet box). 1/enabled, 0/disabled. Defaults to 1.
• bind_forward_address: The overridden DNS server IP address to forward all requests for your own domain to. Defaults to the value of internal_gateway, which is proper in most if not all cases. Your BIND zone is automatically converted from your tinc_peer_name (i.e. lkwco_mane_lixo becomes lkwco.mane.lixo).
• tld: The network-wide TLD to use. Defaults to lixo.

Wireguard options

Keep in mind that Wireguard is presently an auxilliary satellite connection point. Because of this, you shouldn't re-use any IP addresses related to WG. Consider planning a piece (/32, see wg_prefix) of your network (i.e. 172.31.1.z) where z is an unused address (or your ASN number, like 8, 16, so on) and setting that to wg_address. This is generally only desired on "master" or globally-reachable/hosted nodes to construct an overlay VPN. If a node is hosting WireGuard clients, they can send traffic into other nodes but likely won't receive any traffic back unless wg_routes is set so BIRD can export peer addresses into the BGP routing framework.

• wg_enabled: Set to 1 to enable WireGuard. Defaults to 0/disabled.
• wg_key: Required if wg_enabled=1. The private key to use for the Wireguard peering point.
• wg_address: Required if wg_enabled=1. The address to assign to the Wireguard interface. Should be unique.
• wg_prefix: The prefix to assign to Wireguard, defaults to 32.
• wg_port: Port for WireGuard to listen to connections on (UDP). Defaults to 51820.
• wg_routes: A comma-delimited list of WireGuard route prefixes (CIDR) to advertise to other BGP peers. If not specified, no routes are propagated. If specified, only specific connected WireGuard peers added to the kernel routing table are advertised.

IPv6 options (dual-stack)

Enabling IPv6 allows you to be more easily/natively reached by IPv6 native clients such as mobile phones on LTE. Keep in mind this doesn't switch on IPv6 inside the VPN itself. This is typically something you want to do if you are using Wireguard and plan to allow mobile phones to reach you.

• ipv6_enabled: Set to true to enable IPv6 support. Defaults to false.
• internal_address6: Optional; if unset, SLAAC is used. Specify a static IPv6 address to bind to internal_interface.
• internal_mask6: Required if internal_address6 is set. Specify a static IPv6 subnet mask (i.e. 64 for a /64) that is associated to your internal_interface.
• internal_gateway6: Required if internal_address6 is set. Specify a static IPv6 gateway that is reachable at your internal_interface.