Virtual Server BGP Settings
BGP Configuration for RackCorp AS56038
All RackCorp VMs globally are capable of advertising prefixes via BGP at no extra cost. The following instructions relate to customers with virtual servers who wish to advertise their public IP addresses.
1) Preparation:
Submit a support ticket via the support portal with the following information:
- Your AS (if you have one, otherwise we will advertise your IPs under AS56038). We can also accept an AS-SET.
- Any prefixes you wish to advertise. If we should render this via IRR, we can also do this.
Please ensure you add the following to your whois data for your AS (RPSL):
import: from AS56038 accept ANY
export: to AS56038 announce ASXXXXXXXX
where ASXXXXXXXX is your ASN
2) Set up a bgp service on your VM. We recommend using bird, however you can use any daemon you wish. BGP Sessions are made to 110.232.119.251 and 110.232.119.252 regardless of your VM location in the world.
3) Note that it may initially take up to 24 hours for some of our upstreams to pick up and start carrying your prefixes.
Also consider:
RackCorp BGP Communities - LINK
Particularly if you're using Anycast as RackCorp takes great care to try to give a globally balanced anycast network, so consider using 56038:888 community if you're using RackCorp for Anycast BGP services.
Example BGP Configurations
Bird BGP v2 example configuration
Under Centos 8 / Rocky 8 Linux:
yum install epel-release
yum install bird
cat /etc/bird.conf
log syslog all;
router id YOURSERVERIPADDRESS;
protocol device {
scan time 5;
}
protocol direct {
interface "dummy*";
ipv4;
ipv6;
}
protocol bgp rackcorpannounce1
{
local as 56038;
source address YOURSERVERIPADDRESS;
neighbor 110.232.119.251 as 56038;
ipv4 {
import none;
export all;
};
}
protocol bgp rackcorpannounce2
{
local as 56038;
source address YOURSERVERIPADDRESS;
neighbor 110.232.119.252 as 56038;
ipv4 {
import none;
export all;
};
}
protocol static {
ipv4;
# ignore the fact that it says reject, this is where your IP prefixes go
route X.X.X.X/32 reject;
}
No Comments