Accounts 
Locations 
Deployment Options 
Servers 
Operating Systems 
Networking 
Layer 2 Networking 
Equinix Interconnect 
Storage 
Resilience and Recovery 
CLI and Libraries 
Integrations 
COMMAND LINE
Copy to clipboard
Route BGP with BIRD¶
BIRD is an open source routing daemon for Unix-like systems. It can be used to establish bgp sessions between client instances and the Equinix Metal network.
An elastic IP address can simply be announced via BIRD from the instance that is currently using it, thereby making it easy to freely move the IP from one instance to another within the same project.
Getting Started¶
If you haven't already requested that BGP be added to your account you'll need to get that sorted before continuing with this guide - see more info about getting started here.
Choose an IP or Subnet to Advertise¶
Navigate over to IPs & Networks in your BGP enabled project and click on Manage Block for the IPv4 block in the data center location that corresponds with your server deployment. Choose an available IP or subnet that you wish to advertise from your server. In this guide, we'll be using 10.99.200.138/32.
Update Network Interface¶
Update the network interfaces with a virtual loopback interface.
cat >>/etc/network/interfaces <<EOF
auto lo:0
iface lo:0 inet static
address 10.99.200.138
netmask 255.255.255.255
EOF
Bring Up the Interface¶
ifup lo:0
Configuration¶
First install system dependencies. On Ubuntu 18.04 this looks like:
apt -y update && apt -y install python3.6 python3-pip git bird
Now clone the repo and install remaining python dependencies:
cd /opt
git clone https://github.com/packethost/network-helpers.git
cd network-helpers
pip3 install jmespath
pip3 install -e .
Now use the configure.py script to configure BIRD (Legacy Facility example below):
./configure.py -r bird | tee /etc/bird/bird.conf
filter packet_bgp {
# the IP range(s) to announce via BGP from this machine
# these IP addresses need to be bound to the lo interface
# to be reachable; the default behavior is to accept all
# prefixes bound to interface lo
# if net = A.B.C.D/32 then accept;
accept;
}
router id 10.99.182.129;
protocol direct {
interface "lo"; # Restrict network interfaces BIRD works with
}
protocol kernel {
persist; # Don't remove routes on bird shutdown
scan time 20; # Scan kernel routing table every 20 seconds
import all; # Default is import all
export all; # Default is export none
}
# This pseudo-protocol watches all interface up/down events.
protocol device {
scan time 10; # Scan interfaces every 10 seconds
}
protocol bgp neighbor_v4_1 {
export filter packet_bgp;
local as 65000;
neighbor 10.99.182.128 as 65530;
password "somepassword";
}
Below is a bird.conf file generated from the same 'configure.py' script in an IBX. One thing to note is the different BGP neighbor sections to account for the different network infrastructures between IBX'es and Legacy Facilities. Due to this difference, ip routes need to be added to reach the BGP neighbor in all IBX'es, which is reflected in the same IBX bird.conf example below:
./configure.py -r bird | tee /etc/bird/bird.conf
filter packet_bgp {
# the IP range(s) to announce via BGP from this machine
# these IP addresses need to be bound to the lo interface
# to be reachable; the default behavior is to accept all
# prefixes bound to interface lo
# if net = A.B.C.D/32 then accept;
accept;
}
router id 10.68.59.129;
protocol direct {
interface "lo"; # Restrict network interfaces BIRD works with
}
protocol kernel {
persist; # Don't remove routes on bird shutdown
scan time 20; # Scan kernel routing table every 20 seconds
import all; # Default is import all
export all; # Default is export none
}
protocol static {
route 169.254.255.1/32 via 136.144.57.168;
route 169.254.255.2/32 via 136.144.57.168;
}
# This pseudo-protocol watches all interface up/down events.
protocol device {
scan time 10; # Scan interfaces every 10 seconds
}
protocol bgp neighbor_v4_1 {
export filter packet_bgp;
local as 65000;
neighbor 169.254.255.1 as 65530;
password "somepassword";
}
protocol bgp neighbor_v4_2 {
export filter packet_bgp;
local as 65000;
neighbor 169.254.255.2 as 65530;
password "somepassword";
}
Check that the config looks correct, then restart BIRD:
systemctl restart bird
And verify:
birdc
bird> show protocols all neighbor_v4_1
name proto table state since info
neighbor_v4_1 BGP master up 15:20:31 Established
Preference: 100
Input filter: ACCEPT
Output filter: packet_bgp
Routes: 0 imported, 1 exported, 0 preferred
Route change stats: received rejected filtered ignored accepted
Import updates: 0 0 0 0 0
Import withdraws: 0 0 --- 0 0
Export updates: 1 0 0 --- 1
Export withdraws: 0 --- --- --- 0
BGP state: Established
Neighbor address: 10.99.182.128
Neighbor AS: 65530
Neighbor ID: 147.75.36.73
Neighbor caps: refresh restart-aware AS4
Session: external AS4
Source address: 10.99.182.129
Hold timer: 66/90
Keepalive timer: 18/30
In this case we only have a single elastic IP bound to interface to, and we see the prefix is being exported and accepted so we are done.
Note: BIRD uses separate daemons for IPv4 and IPv6 routing. To configure bgp over IPv6 using BIRD, the process will be the same as above, except that the configuration command you will use will be:
./configure.py -r bird6 | tee /etc/bird/bird6.conf
Then:
systemctl restart bird6
In this case we only have a single elastic IP bound to interface to, and we see the prefix is being exported and accepted so we are done.
To test, you can ping the IP address in a command line - ping 10.99.200.138. Remember, Local BGP is announcing a private IP address, so you'll have to be connected to the private network for the data center you're running Local BGP in. You can do that by SSHing into another server in that data center or by connected to Doorman, Equinix Metal's private VPN.
Congratulations! You've just configured a BGP routing protocol.
Alternative Method¶
For those interested, BIRD can also be automatically deployed via docker. See more about deploying BIRD with Docker.