Sending BGP routes using ExaBGP

Hey all,

Recently i came across a problem statement where i wanted to send some good amount of routes via BGP.

While hunting, i came across ExaBGP.

ExaBGP is an open-source software project designed to enable the integration of BGP (Border Gateway Protocol) into various applications and systems.

so, in this post i will try to explain how i am using it.

Topology

If you have gone through my last post on segment routing with Arista, i used a 3 router topology, here i integrated ExaBGP container in the same topology. ExaBGP is spawned in container-lab as an container.

ExaBGP(eth1)---(eth5) R1----R2-----R3
        

ExaBGP has a docker container format, which u can pull from docker registry, but for me when i was trying to integrate with container-lab it was not working, hence i looked how i can make a container myself which can be used in container-lab.

ExaBGP container :

Create folder named exa-image.

create a file named Dockerfile inside it,with below content

run the docker build command to create the docker image.

dshaw@dshaw-pc:~/exa-image$ ls -lrt
total 4
-rw-rw-r-- 1 dshaw dshaw 203 Feb 14 18:15 Dockerfile
dshaw@dshaw-pc:~/exa-image$ cat Dockerfile 
FROM ubuntu:20.04

# Install dependencies
RUN apt update
RUN apt install python3-pip net-tools wget mrtparse vim nano -y && \
    rm -rf /var/lib/apt/lists/* && apt clean
RUN pip install exabgp==4.2.17
        

verify the docker image here.

dshaw@dshaw-pc:~/exa-image$ docker images | grep exa

exa_new   latest   1b3aec26feef   4 weeks ago     486MB         

Container-lab topology :

dshaw@dshaw-pc:~/arsta-seg$ cat topo.yaml 
name: arsta-seg
topology:
  nodes:
    R1:
      kind: ceos
      image: ceos:4.31.2F
    R2:
      kind: ceos
      image: ceos:4.31.2F
    R3:
      kind: ceos
      image: ceos:4.31.2F
    exa1:      
      kind: linux
      image: exa_new:latest
      binds:
        - exa1:/home
  links:
          - endpoints: ["R1:eth1", "R2:eth1"]
          - endpoints: ["R2:eth2", "R3:eth2"]
          - endpoints: ["R1:eth3", "R3:eth3"]
          - endpoints: ["R1:eth5", "exa1:eth1"]        

the bind command here is basically mounting the content of exa1 folder into the docker, in /home directory. Below are the content of the Exa1 directory.

dshaw@dshaw-pc:~/arsta-seg/exa1$ ls -lrt
-rwxrwxr-x 1 dshaw dshaw  46 Feb 14 23:19 deploy.sh
-rw-rw-r-- 1 dshaw dshaw 370 Feb 15 00:06 ebgp-code.conf
-rw-rw-r-- 1 dshaw dshaw 220 Feb 15 00:10 route_gen.py
        

File1: deploy.sh

its optional, i am using it to configure the interface connected to R1. You can simply run the ifconfig command within the ExaBGP container after spinning up the topology

dshaw@dshaw-pc:~/arsta-seg/exa1$ cat deploy.sh 
ifconfig eth1 191.0.0.2 netmask 255.255.255.0
        

File2 : route_gen.py

its a python program that basically controls route generation, you can change the program to generate more routes. I used range 1 to 20, which will generate 19 routes. 10.0.0.1,10.0.0.2 .... like this.

dshaw@dshaw-pc:~/arsta-seg/exa1$ cat route_gen.py 
#!/usr/bin/env python

import sys
import time

for item in range(1,20):
    sys.stdout.write(f'announce route 10.0.0.{item}/32 next-hop 191.0.0.2\n')
    sys.stdout.flush()
    time.sleep(.01)

while True:
	time.sleep(1)
        

File3 : ebgp-code.conf

This file defines the BGP part, and calls the above python program for route generation.

dshaw@dshaw-pc:~/arsta-seg/exa1$ cat ebgp-code.conf 
process gen_routes {
	run ./usr/bin/python3 /home/route_gen.py;
	encoder text;
}

template {
        neighbor leaf1 {
                local-as 101;
                peer-as 100;
                router-id 191.0.0.2;
                local-address 191.0.0.2;
		api {
			processes [ gen_routes ];
		}
                }
        }

neighbor 191.0.0.1 {
        inherit leaf1;
}        

R1 configuration :

here on R1, i am configuring eth5 looking towards ExaBGP container & the required BGP config. I have some left over config from my last lab, hence marked the required ExaBGP things with --->> ExaBGP<<--- marker.

R1#show run int eth5
interface Ethernet5 --->> ExaBGP<<---
   no switchport --->> ExaBGP<<---
   ip address 191.0.0.1/24 --->> ExaBGP<<---
R1#show run sec bgp
router bgp 100
   router-id 1.1.1.1
   neighbor R3 peer group
   neighbor R3 remote-as 100
   neighbor R3 next-hop-self
   neighbor R3 update-source Loopback0
   neighbor exa peer group --->> ExaBGP<<---
   neighbor exa remote-as 101 --->> ExaBGP<<---
   neighbor v6 peer group
   neighbor v6 remote-as 100
   neighbor v6 update-source Loopback0
   neighbor 3.3.3.3 peer group R3
   neighbor 191.0.0.2 peer group exa --->> ExaBGP<<---
   neighbor 9000::3 peer group v6
   !
   address-family ipv4
      neighbor 3.3.3.3 activate
      neighbor 191.0.0.2 activate --->> ExaBGP<<---
      network 5.10.0.0/24
   !
   address-family ipv6
      neighbor 9000::3 activate
      network 5001::/64
        

Starting ExaBGP program:

command1 is to go into the docker container.

command2 and 3 are showing whats mounted.

Command4 starts ExaBGP program.

dshaw@dshaw-pc:~/arsta-seg/exa1$ docker exec -it clab-arsta-seg-exa1 bash
root@exa1:/# cd home/
root@exa1:/home# ls -lrt
-rwxrwxr-x 1 1000 1000  46 Feb 14 17:49 deploy.sh
-rw-rw-r-- 1 1000 1000 370 Feb 14 18:36 ebgp-code.conf
-rw-rw-r-- 1 1000 1000 220 Feb 14 18:40 route_gen.py
root@exa1:/home# exabgp ebgp-code.conf 

if things are fine, then BGP will come up.
18:37:25 | 60     | reactor       | connected to peer-1 with outgoing-10 191.0.0.2-191.0.0.1
        

Validation on R1

R1#show ip bgp summary 
BGP summary information for VRF default
Router identifier 1.1.1.1, local AS number 100
Neighbor Status Codes: m - Under maintenance
  Neighbor  V AS           MsgRcvd   MsgSent  InQ OutQ  Up/Down State   PfxRcd PfxAcc
  3.3.3.3   4 100              255       263    0    0 00:44:58 Estab   1      1
  191.0.0.2 4 101               45        53    0    0 00:41:34 Estab   19     19        

R1#show ip route bgp interface ethernet 5

 B E      10.0.0.1/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.2/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.3/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.4/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.5/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.6/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.7/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.8/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.9/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.10/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.11/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.12/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.13/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.14/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.15/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.16/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.17/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.18/32 [200/0]
           via 191.0.0.2, Ethernet5
 B E      10.0.0.19/32 [200/0]
           via 191.0.0.2, Ethernet5        

so we are receiving the routes from ExaBGP container.

lets go to R3, to see if we are receiving them

R3#show ip route 10.0.0.19/32
 B I      10.0.0.19/32 [200/0]
           via 1.1.1.1/32, IS-IS SR tunnel index 3
              via 1.23.1.2, Ethernet2, label 900041
        

Yes, i am receiving them on R3.

thanks for reading , hope it helps.

While trying to make the ExaBGP container, i faced some issue, but this blog from Julio helped a lot

https://juliopdx.com/2022/02/25/exabgp-in-the-lab/

https://github.com/Exa-Networks/exabgp