Code Disciples

A blog for all things code

Fri 08 May 2020

Deploying a React application on a Linux server

Posted by Abhishek Pednekar in Linux   

In this post, you will deploy a React application created with the create-react-app package. Unlike a React application that is built from scratch, create-react-app does not require us to install packages like babel or webpack. It enables you to get up and running quickly with a basic React application, without a lot of setup work. You can then build on this foundation to develop a full-featured application.

Web deployments on a bare-bones server can sometimes be tricky. This post will go over all the steps that need to be followed to deploy a React application on a virtual private server. The server will be a DigitalOcean droplet. The same steps can be followed on servers provisioned by other providers (like Linode). At the end of this article, you will be familiar with the pre-deployment and deployment tasks for a React application.

Pre-requisites

  1. A local git installation (if you are planning to use the same application as this post)
  2. Node.js to test the application locally
  3. A DigitalOcean account or an account with any IAAS provider
  4. One virtual private server running Ubuntu 18.04, including a non-root sudo-enabled user and a firewall
  5. A React application to deploy

Digital Ocean and the initial setup

You can refer to the Deploying a Flask application on a Linux server - Part I article for details regarding signing up for a DigitalOcean account and performing the initial setup on the server. Although the two-parter article is written for the deployment of a Flask application, part one is generic. One recommended change would be to allow port 3000 instead of 5000 while setting up the firewall (step 5).

Our React application

This post will not go into details about the application that will be deployed. At a high level, you will deploy a single-page React application which displays a random Chuck Norris joke from a public API. Refreshing the page or clicking on the button, displays a new joke. The complete code is accessible via the Github repository. The readme file documents the steps to create a local setup.

"Chuck"

The root folder of the application contains a directory called client. All the React components and modules are present in this directory. Its structure is as follows.

.
└── client
    ├── package-lock.json
    ├── package.json
    ├── public
    │   ├── favicon.jpg
    │   └── index.html
    └── src
        ├── App.css
        ├── App.jsx
        ├── components
        ├── context
        └── index.js

Preparing for deployment

This section deals with a set of pre-deployment tasks. These include,

  1. Creating the production build of the application that will be deployed on the server
  2. Creating a server.js file using Node.js and Express. This file will contain the path to the static assets that were built along with the port information
  3. Test the build locally

To build the project, you will cd into the client directory and run the npm run build command. This will create a folder named build inside the client directory. The build folder contains the static assets which will be deployed on the server.

cd client

npm run build

The next step is to a create directory called server in the project root folder (in the same level as the client directory). Within server, you will create a server.js file with the following code.

// server.js

const express = require("express");
const path = require("path");

const app = express();

app.use(express.static(path.join(__dirname, "../client/build")));
app.get("/*", (req, res) => res.sendFile(path.join(__dirname, "./index.html")));

const PORT = process.env.PORT || 3000;

app.listen(PORT, () => console.log(`Server started on port ${PORT}`));

The code in server.js uses Node.js and express to specify the path to the static assets in the build folder. Also, you are ensuring that when accessing the root of the site, the index.html file is presented to an end-user. The application is being configured to listen for requests on port 3000.

To run server.js you need the express package. To do that, you will first cd into the server directory and run npm install express. Finally, the build can be tested by running it locally using the node server.js command. If all goes fine, the application will be accessible on http://localhost:3000.

cd server

npm install express

node server.js

Deploying the application

With the server now primed for deployment and the build tested locally, you can now copy over the necessary files/folders from the local machine to the DigitalOcean droplet. There are several ways to accomplish this. One way is to use the scp command.

Below is the complete list of files/folders that will be copied to the server.

.
├── client
│   ├── build
│   ├── package-lock.json
│   └── package.json
└── server
    └── server.js

On the server, these files/folders will be copied to a folder named react-demo-app located in the home directory of the user demo. You can place these anywhere. Remember to change the username, IP address and path when running the below scp command. Assuming we are in the project's root directory on our local machine (/mnt/c/react-demo-app/ in this case), the command will copy the client and server subdirectories to the server.

cd /mnt/c/react-demo-app/

scp -r * demo@139.59.4.54:~/react-demo-app/

Most of the next steps will need to be carried out on our DigitalOcean droplet. You will need to log in with the sudo user to perform the next steps.

ssh demo@@139.59.4.54

Installing Node.js

To install the latest version of Node.js and npm on the server, the following commands should be run.

sudo apt-get update

curl -sL https://deb.nodesource.com/setup_12.x | sudo -E bash -

sudo apt install nodejs -y

To verify if the installation was successful, you can run

node --version

npm --version

Installing project dependencies on the server

With Node.js and npm installed, you can now install our project dependencies. First, navigate to the client directory and install the dependencies using the npm install command.

cd ~/react-demo-app/client

npm install

Remember, that the server directory in ~/react-demo-app only contains the server.js file. Also, to run the application using Node.js, you need the express package. To install express, navigate to the server directory and run npm install express.

cd ~/react-demo-app/server

npm install express

At this time, you can do a quick test to see if the site works by running node server.js. The site should be accessible at http://<ipaddress>:3000. Port 3000 will work provided the firewall (ufw) was configured to allow it.

cd ~/react-demo-app/server

node server.js


site

Almost there! Disconnecting from the terminal will kill the process that is running the site. To keep the site running at all times, you will next install the PM2 process manager.

Installing PM2 to run the site

PM2 will be installed as a global dependency using the sudo npm install -g pm2 command. Once installed PM2 can be used to start and manage the site via the pm2 start command.

// Install PM2
sudo npm install -g pm2

// Navigate to the server directory and run the site with PM2
cd ~/react-demo-app/server

pm2 start server.js

You should see an output similar to this -

pm2

Now, the site will continue to be accessible on http://<ipaddress>:3000 despite disconnecting from the terminal. To check the status of PM2, you can run sudo pm2 status. You, however, don't want your users to always specify a port number while accessing the site. To ensure this, incoming traffic will be routed through the Nginx web server.

Installing and configuring Nginx

Nginx will act as the webserver and route incoming traffic based on the specified configuration. You will need to first install it on the server and add some configuration.

// Installing Nginx
sudo apt install nginx -y

Before creating a configuration file for Nginx, you will need to make sure that Nginx can find the configuration that you are about to create. Check the contents of the nginx.conf file located in /etc/nginx using cat /etc/nginx/nginx.conf and look for the line include /etc/nginx/sites-enabled/*; in the http directive. If it does not exist, you will need to add it.

To add the line, open nginx.conf using the nano (or any suitable) editor.

sudo nano /etc/nginx/nginx.conf

http {
    ...
    ...

    include /etc/nginx/sites-enabled/*;
}

Next, delete the default nginx configuration and create a new configuration file called demo (it can be named anything) in the /etc/nginx/sites-enabled/ directory.

// Delete existing Nginx configuration
sudo rm /etc/nginx/sites-enabled/default

sudo nano /etc/nginx/sites-enabled/demo

Add the below code to the demo file.

    server_name 139.59.4.54; # Whatever is your IP or domain

    location / {
        proxy_pass http://localhost:3000; # Whatever port your app runs on
        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection 'upgrade';
        proxy_set_header Host $host;
        proxy_cache_bypass $http_upgrade;
    }

To check the syntax of the nginx configuration file, you can run sudo nginx -t.

This is a very basic configuration and will suffice for this demo project. Based on this current configuration, Nginx will forward any request that comes to the root / of the site to http://localhost:3000. Remember that PM2 is running our site on port 3000 on the server. This way, users will just need to enter the domain name or IP address in their browsers and Nginx will make sure that their requests are routed correctly.

The proxy_set_header directives in the configuration will be included in the request headers. This is by no means the most complete Nginx configuration but a bare minimum one that serves the current purpose. The Nginx documentation lists the complete set of directives.

Once the configuration file is saved, you can restart Nginx.

// Restarting Nginx

sudo systemctl restart nginx

The last step will be to configure the firewall to allow traffic on the http/tcp port. At this point, you can also delete port 3000 from our firewall rules.

// Add http
sudo ufw allow http/tcp

// Delete 3000
sudo ufw delete allow 3000

// Restart ufw
sudo ufw enable

This is what the status of our firewall should now look like.

Status: active

To                         Action      From
--                         ------      ----
22/tcp                     ALLOW       Anywhere
80/tcp                     ALLOW       Anywhere
22/tcp (v6)                ALLOW       Anywhere (v6)
80/tcp (v6)                ALLOW       Anywhere (v6)

That's it! The fully functioning site will now be accessible on http://<ipaddress>.

site2

Next steps

As next steps, you can create a CI/CD pipeline for automated deployments. Please refer to the Continuous Deployment with Travis CI and DigitalOcean post for more details.


Share



Comments