Layer 4 load balancing using Nginx

In this tutorial, I am going to demonstrate Layer 4 load balancing using Nginx.

What we will learn from here?

• VPC Creating

• VPC peering

• use of docker-compose
• Writing Nginx configuration as load balancer, and many more commands and uses of different tools

---

Let's get started

Firstly, see the diagram of what we are going to do.

Our plan is to receive requests into Nginx if it will in port 80 then it will forward it to the web server or our application server. If the application server want to access the database it will send a request to Nginx on port 3306. If Nginx received any request into the 3306 port, so it will forward the request to the MySQL server. We will create 3 VPCs and also create a VM inside each. So each VPC has a VM and set IP range for each as below.

• VPC 1

  • VM: for Nginx
  • IP Range : 10.1.0.0/16

• VPC 2

  • VM: for Python / PHP
  • IP Range : 10.2.0.0/16

• VPC 3

  • VM: for MySQL
  • IP Range : 10.3.0.0/16

---

• The 1st VM will contain Nginx as a load balancer.
• The 2nd VM will contain a Web server (or Application Server), we can install PHP or python.
• The last one will contain MySQL as the Database server.

Now create VPC peering from both sides as below

VPC 1 to VPC 2
VPC 2 to VPC 1

VPC 1 to VPC 3
VPC 3 to VPC 1

After successful peering, it will show active status.

Let's create VM in each VPC Then access into each.

You can see 3 instances here. Click SSH on Nginx instance, it will open on a new tab but will not connect. because SSH use Google IAP service to get access into VM. So we need to allow traffic from 35.235.240.0/20 . To fix the issue you need to create a firewall rule and allow 35.235.240.0/20 for all ingress traffic.

To know more visit here : Google IAP

After retry by clicking again on SSH and hope you will get access into nginx instance.

• Install Docker
• Install docker-compose
• Create docker-compose.yml file into it by sudo nano docker-compose.yml

Paste the following code

  version: "3"
  services:
  load-balance:
  image: nginx:stable-alpine
  container_name: nginx-proxy
  ports:
  - "80:80"
  volumes:
  - ./prod/nginx.conf:/etc/nginx/ngix.conf

Create nginx.conf file by sudo nano nginx.conf Paste the following code

stream {
        upstream php-server {
        server 10.2.0.0:80; # private ip of php vpc
        }
        
        upstream mysql-db {
        server 10.3.0.0:3306; # private ip of mysql vpc
        }
        
    server {
        listen 80;
        proxy_connect_timeout 5s;
        proxy_timeout 5s;
        proxy_pass php-server;
        }
        
    server {
        listen 3306;
        proxy_connect_timeout 5s;
        proxy_timeout 5s;
        proxy_pass mysql-db;
        }
    }

Run the container by docker-compose up -d (-d for running into detached mode -i.e: in background)

To check running container you can run docker ps (ps-process status)

• Do the same process for others. Create docker-compose.yml for mysql

for the python also. (check the folder above)

The load balancer (nginx) is configured to listen on port 80 and forward to the Application service on another VPC (10.2.0.2) port 3000. And port 3306 is forward to the database server instance's , i.e: 10.3.0.3 to 3306 port that also on another VPC's .

Test 01: Nginx to web server

To test, from menu go to Compute engine > VM instance >

• Click SSH on Nginx instance and get access into it.

• Install telnet by running the command:sudo apt install telnet

• Then run command to check webapp connection telnet 10.2.0.2 3000

• before that create firewall rule to allow nginx IP 10.1.0.2 into webserver as ingress request

Test 02: Nginx to MySQL

• create firewall rule to allow nginx IP 10.1.0.2 into MySQL as ingress request

• To check database connection telnet 10.3.0.2 3306

• To check database connection telnet 10.0.0.2 3306

Final test: web server to mysql

• create firewall rule to allow MySQL IP 10.3.0.2 into Nginx as ingress request
• To check database connection telnet 10.1.0.2 3306 If everything is fine i.e: webserver send the request to the nginx and nginx forward the request to the mysql.