Deploy React App To Heroku Using Postgres & Express
Deploy React App To Heroku Using Postgres & Express
Heroku is an web application that makes deploying applications easy for a beginner.
# ############################################################################## | |
##### HEROKU TOOLBELT COMPLETE GUIDE ########################################### | |
################################################################################ | |
# Installing Heroku toolbelt using command line | |
# For MacOS... | |
brew tap heroku/brew && brew install heroku | |
# For Ubuntu... | |
sudo snap install --classic heroku | |
# Other installation methods are | |
curl https://cli-assets.heroku.com/install.sh | sh # only for unix based systems, windows incompatible as it needs sudo | |
curl https://cli-assets.heroku.com/install-ubuntu.sh | sh # Ubuntu/Debian apt-get | |
yay -S heroku-cli # Arch linux, Note: This package is community maintained not by heroku | |
npm install -g heroku # This installation method is required for users on ARM and BSD... | |
############ | |
# Verifying your installation | |
heroku --version | |
# Let's get started with heroku | |
heroku login # To login into the heroku toolbelt with your heroku account, this will open browser for you. | |
heroku login -i # If you prefer to stay in the command line environment, then you can execute this command | |
# Now navigate to your desired directory and create a blank heroku application | |
cd ~/myapp | |
heorku create | |
# If you are facing login issues, try to execute the following command | |
mv ~/.netrc ~/.netrc.backup | |
heroku login | |
# Uninstalling the heroku CLI | |
# For macOS | |
rm -rf /usr/local/heroku /usr/local/lib/heroku /usr/local/bin/heroku ~/.local/share/heroku ~/Library/Caches/heroku | |
# or you can try the below command also on macOS | |
brew uninstall heroku | |
rm -rf ~/.local/share/heroku ~/Library/Caches/heroku | |
# For Linux (Standalone installs) | |
rm /usr/local/bin/heroku | |
rm -rf /usr/local/lib/heroku /usr/local/heroku | |
rm -rf ~/.local/share/heroku ~/.cache/heroku | |
# For Linux (Debian and Ubuntu installs) | |
sudo apt-get remove heroku heroku-toolbelt | |
sudo rm /etc/apt/sources.list.d/heroku.list | |
##################################################################################################### | |
### Managing and deploying applications on Heroku (Using Git) ################################### | |
##################################################################################################### | |
cd myapp # Changing into the project directory | |
git init # Initializing the project into a git repository | |
git add -f example.json # Adding a perticular content of the project into the repository this will include the content from .gitignore | |
git add . # Adding all the contents of the project into the repository excluding .gitignore content | |
git commit -m "My first commit" # Commiting the content to the repository | |
heroku create appname # Creating a new application on Heroku here ( appname ) represent the name u give to your app | |
git remote -v # verifying that the remote is set to the heroku | |
heroku git:remote -a thawing-inlet-61413 # For an existing heroku app, you can add remote to the application | |
git remote rename heroku heroku-staging # renaming remotes | |
git push heroku master # Deploying code to the heroku application | |
git push heroku master --force # Force Pushing to heroku ( required if the remote contain works that u do not have locally ) | |
git push heroku testbranch:master # Deploying code from a non-master branch to the heroku application | |
heroku create --ssh-git # ssh git transport for the application instead of https | |
git config --global url.ssh://git@heroku.com/.insteadOf https://git.heroku.com/ # For using ssh always | |
git config --global --remove-section url.ssh://git@heroku.com/ # To remove this rewrite setting run the command | |
##################################################################################################### | |
### Managing and deploying applications on Heroku (Using Docker) ################################### | |
##################################################################################################### | |
# Setting stack of your app to a Container | |
heroku stack:set container | |
heroku container:login # Login to the container resistry | |
git clone https://github.com/heroku/alpinehelloworld.git # Get sample code by cloning into the following repository | |
heroku create appname # Creating a heroku application here ( appname ) represent the name u give to your app | |
heroku container:push web # Build the image and push to Container Registry | |
heroku container:push --recursive # Pushing from the root directory of the project in recursive manner | |
heroku container:push web worker --recursive # Building the image and pushing to container resistry in recursive manner | |
heroku container:release web # Releasing the image to your application | |
heroku open # Open the application in the browser | |
Before you begin deploying, make sure to remove any
console.log
's or
debugger
's
in any production code. You can search your entire project
folder if you are using them anywhere.
You will set up Heroku to run on a production, not development,
version of your application. When a Node.js application like
yours is pushed up to Heroku, it is identified as a Node.js
application because of the
package.json
file. It runs
npm install
automatically. Then, if there is a
heroku-postbuild
script in the
package.json
file, it will run that script. Afterwards, it will automatically
run npm start
.
In the following phases, you will configure your application to
work in production, not just in development, and configure the
package.json
scripts for
install
,
heroku-postbuild
and
start
scripts
to install, build your React application, and start the Express
production server.
Phase 1: Heroku Connection
If you haven’t created a Heroku account yet, create one here.
Add a new application in your Heroku dashboard named whatever you want. Under the “Resources” tab in your new application, click “Find more add-ons” and add the “Heroku Postgres” add-on with the free Hobby Dev setting.
In your terminal, install the Heroku CLI. Afterwards, login to Heroku in your terminal by running the following:
heroku login
Add Heroku as a remote to your project’s git repository in the
following command and replace
<name-of-Heroku-app>
with the name of the application you created in the
Heroku dashboard.
heroku git:remote -a <name-of-Heroku-app>
Next, you will set up your Express + React application to be deployable to Heroku.
Phase 2: Setting up your Express + React application
Right now, your React application is on a different localhost
port than your Express application. However, since your React
application only consists of static files that don’t need to
bundled continuously with changes in production, your Express
application can serve the React assets in production too. These
static files live in the
frontend/build
folder after running
npm run build
in the
frontend
folder.
Add the following changes into your
backend/routes.index.js
file.
At the root route, serve the React application’s static
index.html
file
along with
XSRF-TOKEN
cookie. Then serve up all the React application's static
files using the
express.static
middleware. Serve the
index.html
and
set the
XSRF-TOKEN
cookie again on all routes that don't start in
/api
. You
should already have this set up in
backend/routes/index.js
which should now look like this:
// backend/routes/index.js
const express = require('express');
const router = express.Router();
const apiRouter = require('./api');
router.use('/api', apiRouter);
// Static routes
// Serve React build files in production
if (process.env.NODE_ENV === 'production') {
const path = require('path');
// Serve the frontend's index.html file at the root route
router.get('/', (req, res) => {
res.cookie('XSRF-TOKEN', req.csrfToken());
res.sendFile(
path.resolve(__dirname, '../../frontend', 'build', 'index.html')
);
});
// Serve the static assets in the frontend's build folder
router.use(express.static(path.resolve("../frontend/build")));
// Serve the frontend's index.html file at all other routes NOT starting with /api
router.get(/^(?!\/?api).*/, (req, res) => {
res.cookie('XSRF-TOKEN', req.csrfToken());
res.sendFile(
path.resolve(__dirname, '../../frontend', 'build', 'index.html')
);
});
}
// Add a XSRF-TOKEN cookie in development
if (process.env.NODE_ENV !== 'production') {
router.get('/api/csrf/restore', (req, res) => {
res.cookie('XSRF-TOKEN', req.csrfToken());
res.status(201).json({});
});
}
module.exports = router;
Your Express backend’s
package.json
should include scripts to run the
sequelize
CLI
commands.
The
backend/package.json
's scripts should now look like this:
"scripts": {
"sequelize": "sequelize",
"sequelize-cli": "sequelize-cli",
"start": "per-env",
"start:development": "nodemon -r dotenv/config ./bin/www",
"start:production": "node ./bin/www"
},
Initialize a
package.json
file at the very root of your project directory (outside of both
the backend
and
frontend
folders). The scripts defined in this
package.json
file will be run by Heroku, not the scripts defined in the
backend/package.json
or the
frontend/package.json
.
When Heroku runs
npm install
, it
should install packages for both the
backend
and the
frontend
.
Overwrite the
install
script
in the root
package.json
with:
npm --prefix backend install backend && npm --prefix frontend install frontend
This will run
npm install
in
the
backend
folder
then run
npm install
in
the
frontend
folder.
Next, define a
heroku-postbuild
script that will run the
npm run build
command in the
frontend
folder. Remember, Heroku will automatically run this script
after running
npm install
.
Define a
sequelize
script that will run
npm run sequelize
in the
backend
folder.
Finally, define a
start
that will
run
npm start
in
the `backend folder.
The root
package.json
's scripts should look like this:
"scripts": {
"heroku-postbuild": "npm run build --prefix frontend",
"install": "npm --prefix backend install backend && npm --prefix frontend install frontend",
"dev:backend": "npm install --prefix backend start",
"dev:frontend": "npm install --prefix frontend start",
"sequelize": "npm run --prefix backend sequelize",
"sequelize-cli": "npm run --prefix backend sequelize-cli",
"start": "npm start --prefix backend"
},
The
dev:backend
and
dev:frontend
scripts are optional and will not be used for Heroku.
Finally, commit your changes.
Phase 3: Deploy to Heroku
Once you’re finished setting this up, navigate to your
application’s Heroku dashboard. Under “Settings” there is a
section for “Config Vars”. Click the
Reveal Config Vars
button to see all your production environment variables. You
should have a
DATABASE_URL
environment variable already from the Heroku Postgres add-on.
Add environment variables for
JWT_EXPIRES_IN
and
JWT_SECRET
and
any other environment variables you need for production.
You can also set environment variables through the Heroku CLI you installed earlier in your terminal. See the docs for Setting Heroku Config Variables.
Push your project to Heroku. Heroku only allows the
master
branch
to be pushed. But, you can alias your branch to be named
master
when
pushing to Heroku. For example, to push a branch called
login-branch
to
master
run:
git push heroku login-branch:master
If you do want to push the
master
branch,
just run:
git push heroku master
You may want to make two applications on Heroku, the
master
branch
site that should have working code only. And your
staging
site
that you can use to test your work in progress code.
Now you need to migrate and seed your production database.
Using the Heroku CLI, you can run commands inside of your
production application just like in development using the
heroku run
command.
For example to migrate the production database, run:
heroku run npm run sequelize db:migrate
To seed the production database, run:
heroku run npm run sequelize db:seed:all
Note: You can interact with your database this way as you’d
like, but beware that
db:drop
cannot
be run in the Heroku environment. If you want to drop and create
the database, you need to remove and add back the "Heroku
Postgres" add-on.
Another way to interact with the production application is by opening a bash shell through your terminal by running:
heroku bash
In the opened shell, you can run things like
npm run sequelize db:migrate
.
Open your deployed site and check to see if you successfully deployed your Express + React application to Heroku!
If you see an
Application Error
or are experiencing different behavior than what you see in your
local environment, check the logs by running:
heroku logs
If you want to open a connection to the logs to continuously output to your terminal, then run:
heroku logs --tail
The logs may clue you into why you are experiencing errors or different behavior.
If you found this guide helpful feel free to checkout my github/gists where I host similar content:
Web Developer, Electrical Engineer JavaScript | CSS | Bootstrap | Python | React | Node.js | Express | Sequelize…github.com
Alternate Instructions:
You've got a React app, and an API server written in Express or something else. Now -- how do you deploy them both to a server?
There are a few ways to do this:
- Keep them together -- Express and React files sit on the same machine, and Express does double duty: it serves the React files, and it also serves API requests.
- e.g., a DigitalOcean VPS running Express on port 80
- Split them apart -- Host the Express API on one machine, and the React app on another.
- e.g., React app served by Amazon S3, API server running on a DigitalOcean VPS
- Put the API behind a proxy -- Express and React app files sit on the same machine, but served by different servers
- e.g., NGINX webserver proxies API requests to the API server, and also serves React static files
This article will cover how to keep them together. We'll build the Express server to serve React's static files in addition to providing an API, and then deploy it to Heroku. Heroku is easy to deploy to and free to get started with.
If you don't have one already, go here and sign up. It's free.
Heroku comes with a commandline command they call a "toolbelt." Follow the instructions here to install it. (On a Mac with Homebrew, just brew install heroku
).
We'll build a password generator. Every time you load the app or click Get More you'll get 5 random paswords.
Just a quick disclaimer: this is just meant as a demo! I don't recommend using some random internet thing that generates passwords on the server to generate your own real passwords ;)
Make a parent directory to contain everything. Call it rando
or whatever you want.
$ mkdir rando; cd rando
Initialize the project with Yarn or NPM:
$ yarn init -y
# or npm init -y
We need 2 packages: Express itself, and a password generator. Install those now:
$ yarn add express password-generator
Create a file called index.js
, which will be the Express app, and type this in:
const express = require('express');
const path = require('path');
const generatePassword = require('password-generator');
const app = express();
// Serve static files from the React app
app.use(express.static(path.join(__dirname, 'client/build')));
// Put all API endpoints under '/api'
app.get('/api/passwords', (req, res) => {
const count = 5;
// Generate some passwords
const passwords = Array.from(Array(count).keys()).map(i =>
generatePassword(12, false)
)
// Return them as json
res.json(passwords);
console.log(`Sent ${count} passwords`);
});
// The "catchall" handler: for any request that doesn't
// match one above, send back React's index.html file.
app.get('*', (req, res) => {
res.sendFile(path.join(__dirname+'/client/build/index.html'));
});
const port = process.env.PORT || 5000;
app.listen(port);
console.log(`Password generator listening on ${port}`);
We're also going to need a "start" script in package.json, so that Heroku knows how to start the app. Open package.json
and add a scripts section at the bottom. The full file should look something like this:
{
"name": "rando",
"version": "1.0.0",
"main": "index.js",
"license": "MIT",
"dependencies": {
"express": "^4.15.3",
"password-generator": "^2.1.0"
},
"scripts": {
"start": "node index.js"
}
}
It's always good to make sure things are working as you go along. Much better than getting to the end and realizing nothing works. So, let's try it out.
Start up the Express app by running:
$ yarn start
Open up your browser and go to http://localhost:5000/api/passwords. You should see something like this:
Now we'll deploy the app to Heroku, make sure it works, and then we'll add React to the mix.
Heroku needs your project to have a Git repository, so we'll create one along with a .gitignore
file to ignore node_modules, and then commit the code:
$ git init
$ echo node_modules > .gitignore
$ git add .
$ git commit -m "Initial commit"
Now we're ready for Heroku. Run its 'create' command:
$ heroku create
And you'll see something like this:
To make it work, we just need to push up our code by running:
$ git push heroku master
It will print out a bunch of stuff, and then the app will be live. One of the last lines will tell you the URL of the app:
Now you can go to <your url>/api/passwords
and make sure it works.
Woohoo! You've got an app live on the real internet! Except it's not very nice to use, yet. Let's add a React frontend now.
We're going to use Create React App to generate a project. Remember that we decided the React app would live in the "client" folder? (we did, back when we set up Express to point to "client/build" for static assets).
If you don't have Create React App installed yet, do that first:
$ yarn global add create-react-app
# or npm install -g create-react-app
Generate the React app inside the Express app directory:
$ create-react-app client
Create React App will proxy API requests from the React app to the Express app if we add a "proxy" key in package.json like this:
"proxy": "http://localhost:5000"
This goes in client/package.json
, not in the Express app's package.json, and it will be ignored by Heroku after deploying.
Open up src/App.js
and replace it with this:
import React, { Component } from 'react';
import './App.css';
class App extends Component {
// Initialize state
state = { passwords: [] }
// Fetch passwords after first mount
componentDidMount() {
this.getPasswords();
}
getPasswords = () => {
// Get the passwords and store them in state
fetch('/api/passwords')
.then(res => res.json())
.then(passwords => this.setState({ passwords }));
}
render() {
const { passwords } = this.state;
return (
<div className="App">
{/* Render the passwords if we have them */}
{passwords.length ? (
<div>
<h1>5 Passwords.</h1>
<ul className="passwords">
{/*
Generally it's bad to use "index" as a key.
It's ok for this example because there will always
be the same number of passwords, and they never
change positions in the array.
*/}
{passwords.map((password, index) =>
<li key={index}>
{password}
</li>
)}
</ul>
<button
className="more"
onClick={this.getPasswords}>
Get More
</button>
</div>
) : (
// Render a helpful message otherwise
<div>
<h1>No passwords :(</h1>
<button
className="more"
onClick={this.getPasswords}>
Try Again?
</button>
</div>
)}
</div>
);
}
}
export default App;
You can update the CSS too, if you like (in src/App.css
):
.App {
text-align: center;
font-family: "Courier New", monospace;
width: 100%;
}
h1 {
font-weight: normal;
font-size: 42px;
}
.passwords {
list-style: none;
padding: 0;
font-size: 32px;
margin-bottom: 2em;
}
.more {
font-size: 32px;
font-family: "Courier New", monospace;
border: 2px solid #000;
background-color: #fff;
padding: 10px 25px;
}
.more:hover {
background-color: #FDD836;
}
.more:active {
background-color: #FFEFA9;
}
I also recommend opening up src/index.js
and removing the call to registerServiceWorker()
at the bottom, since it can cause some confusing caching issues (like preventing you from accessing the API endpoints in a browser after you load the React app once).
Now start up the React app by running yarn start
inside the client
folder.
Make sure the Express app is running too: run yarn start
from its folder as well.
Go to http://localhost:3000 and the app should be working! Now we can deploy the whole thing to Heroku.
When you deploy the app with the git push heroku master
command, git copies all the checked-in files up to Heroku. There are two complications now:
- We need to check in the new
client
code - Express depends on the built client code in
client/build
, which we don't have yet, and which we'd rather not check in to git.
What we'll do is tell Heroku to build the React app after we push up the code, and we can do that by adding a "heroku-postbuild" script in the top-level (Express app's) package.json.
If you're using Yarn, the script looks like this:
"scripts": {
"start": "node index.js",
"heroku-postbuild": "cd client && yarn && yarn run build"
}
This tells Heroku "hey, after you're done doing what you do, go into the client folder and build my React app." The yarn run build
script will kick off Create React App's production build, which will put its output files in the client/build
folder so Express can find them.
If you're using NPM, the script will look like this:
"scripts": {
"start": "node index.js",
"heroku-postbuild": "cd client && npm install && npm run build"
}
This tells Heroku "hey, after you're done doing what you do, go into the client folder and build my React app." The npm run build
script will kick off Create React App's production build, which will put its output files in the client/build
folder so Express can find them.
Thanks to Matthew Locke and Babajide Ibiayo in the comments for how to get this working with NPM.
Once you have configured the heroku-postbuild
step for Yarn (or NPM), add everything to git and commit it. Make sure to run this from the top-level rando
directory, not inside client
:
$ git add .
$ git commit -m "Ready for awesome"
If you run git status
now, you should see no red items.
Then you can deploy the app by running:
$ git push heroku master
It again prints out your app's hostname. Mine is https://glacial-brook-33351.herokuapp.com/. Go there and try it out!
Congrats, you've got a React + Express app in production ;)
The complete app can be found on Github, and the README there explains how to deploy it.
Check out the npm
branch with git checkout npm
if you want to use NPM. From there, the deploy will differ slightly -- run git push heroku npm:master
to deploy the npm branch insead of master.
Deploy MERN App To Heroku:
Source: Article
This post is intended to be a guide for those that want to deploy a full-stack MERN app. It will be v...
Nick Aug 29, 2020 ・Updated on Aug 31, 2020 ・12 min read
This post is intended to be a guide for those that want to deploy a full-stack MERN app. It will be very specific on steps so it's a bit of a read, however, it will ensure there is minimal to no confusion on how to get from point A to point B.
Feel free to hop around if you've got some of these steps done, know how to do them already, what have you.
If this is your first time don't intimidate yourself! Just read and complete each step one at a time.
NOTE: This guide is based on my specific project structure and this works well for it. There are definitely other ways to deploy and other ways people prefer to have their project structured.
So you've got a cool project you'd like to show off to the world, how do we deploy a full-stack MERN app?
Let's first get some basics out of the way and context on how I did this: my project structure, basic configuration, and why I chose this way of deployment.
MERN stands for MongoDB - Express - React - Node.
- MongoDB for the database
- Node & Express for the server-side
- React for the client-side
There's also the MEAN stack, which uses Angular instead of React, and the... MEVN(?) stack... whatever, it uses Vue instead of React or Angular.
One thing that is definitely dev to dev, company to company, is how we structure our project as a whole. By this I mean where our directories are in relation to each other. Here is my fully collapsed root directory to show you the top-level:
Just three things:
- My VS Code settings and extensions JSON files
- Client-Side Code
- Server-Side Code
I like this structure because to me there's a clear separation of responsibilities. It makes it easier for me to work on one or the other by just collapsing that whole folder. For instance if there's some UI feature I want to add/fix/alter then I have no reason to open my server
directory.
I feel it makes it logical and simple to work with. Again, this is absolutely personal preference. This is also a big reason why this project was deployed in the way this guide will describe.
Credit to Dave Ceddia for putting this list together - this was found in a blog post of his.
- Keep
client
andserver
together- Your Express and React files will live on the same machine and Express will both serve your API requests and the React files as well
- Put your API behind a proxy
- This will allow your Express and React files to still live on one machine but each will be served by a different server
- Split
client
andserver
apart (the option for this guide)- Host your Express API on one machine and the React App on another
Deployment is not something I have a ton of experience with so the first two options I've never attempted.
I do know that in some cases for deployment on one machine a lot of devs must nest their entire client
directory inside of server
. This was a big reason I went the route I did for deployment because for me I wanted to keep the structure pictured above with a clear separation of client-side code and server-side code.
The base of the project I deployed was the Create-React-App, which comes with a webpack.config.js
and other basic configurations out of the box, for the front-end.
I also used Webpack for the server
directory as well to allow all those beautiful import
s and export
s to work for me and allow me to separate responsibilities on the server-side.
Here's a shot of the client
and server
directories expanded just to show you how I have it setup. There's only a couple things that will need to be there for this project structure to work which I'll point out.
Each directory, client
and server
, needs:
.gitignore
- To ignore the files and directories we don't want stored in our repo
package.json
- To specify our separate
dependencies
anddevDependencies
- To specify our separate
webpack.config.js
- The
client
doesn't have one because Create-React-App again provides one out of the box so I didn't need to specify it there
- The
You may have noticed I have a yarn.lock
inside client
and a package-lock.json
inside server
. I intentionally use yarn
for client-side and npm
for server-side because it helps me just keep those separate in my mind. I know if I'm using yarn
then I'm in client
and if I'm using npm
I'm in server
. Once again just a personal preference - you could use yarn
or npm
for both and have no issues.
You may have also noticed the Procfile
which we'll get to later in the guide. As a teaser this will be needed by Heroku to deploy our server
code which is why it lives inside of server
.
Now that we have some context on how my project is setup let's talk about the actual process of deploying.
What we're going to do specifically is host our server
code on Heroku and our client
code on Netlify. So our React App hosted on Netlify will make API requests to our Express API hosted on Heroku.
This will assume that you have both client
and server
running correctly and that you have already connected your app to a hosted database on MongoDB.
-
Create a Netlify account or sign in if you have one
-
Create a Heroku account or sign in if you have one
-
- There is instructions on how to do this for both Mac and Windows in this linked article
-
Go to your terminal and type
heroku login
- This will prompt you to press any key, once you do it will take you to your browser where you will just need to click 'Log In'.
- Once this is successful you can close that browser window and navigate to your text editor now with your project open
-
Create a file named
Procfile
, no file extension, in/server
directory.- Include this text and this text only in the
Procfile
web: node dist/main.js
- This tells Heroku what to do after building our app - which specifically is to run our bundled file from webpack with all of our server logic inside it
- If we didn't include this it would likely build just fine but never actually start our server
- Include this text and this text only in the
-
Now we must add to our
server/package.json
the following block:-
You can more than likely add this anywhere (not nested in any other property) but I added it right below the
"main"
property near the top in theserver/package.json
. -
Recommended to check your node and npm versions first to know which versions to put in this next block. While it likely won't break your app with a slightly different version it is safer to specify the versions you used in development to prevent unexpected behavior.
-
node -v
ornode --version
to check your Node version -
npm -v
ornpm --version
to check your npm version -
Add those versions that return to this
engines
property if they're different than mine. -
I used
"14.x"
fornode
to say, "As long as the major version of Node is 14 then use the latest minor version. Major version releases are when they're likely to be breaking changes which we want to avoid."engines": { "node": "14.x", "npm": "6.14.7" },
-
-
We need to allow access to your MongoDB database now from a new IP address. For simplicity, I added all IP addresses to be allowed.
-
In development I had the express server listen to
localhost:5000
to run there and my react app ran on port3000
. We must change what this Express server will listen for toapp.listen(process.env.PORT || 5000)
-
So this will tell the server to listen to the
.env
environment variablePORT
which we should not have set in our own.env
file. OR it will listen to port 5000. -
The
process.env.PORT
will be set by Heroku's.env
- they essentially will tell our server which port to listen to. -
If there is no such
process.env.PORT
, in other words there is no.env
file that has been configured, then it will default to5000
. This will allow us to keep that code in and be able to run it in our local development environment on port5000
because we do not have a.env
variablePORT
.// server.js
/* Other code... */
app.listen(process.env.PORT || 3000);
Ensure in your terminal that you are inside the root of the project you are deploying for these next two steps
-
-
Now back inside your Terminal the next command we type is
heroku git:remote -a <project-name>
. Replace<project-name>
with really whatever name you want - I recommend the name of your project to keep it simple. So for example it will look likeheroku git:remote -a my-project
.-
This is going to create a remote connection with Heroku specifically and also create a branch on that remote for you and your app name will be the name you provided, such as
my-project
above. Next we're going to actuallypush
our server code to that remote Heroku branch.heroku git:remote -a my-project
-
-
The command we need to push our
server
code specifically, because we have the separation ofclient
andserver
in our project structure, isgit subtree push --prefix server heroku master
* This is telling git to push a subtree, or in other words a _subdirectory_, from our current git repo. The `--prefix` flag comes before that subdirectory that we want to push, in our case it is `server`. And lastly we're telling it to push to the remote `heroku` our `master` branch.
git subtree push --prefix server heroku master
* * *
**IF you want to test and ensure that you've deployed your server to Heroku just add in a test route in your `server.js` like:**
app.get('/', (req, res) => { res.send('Hello from Express!')
You can view your app by:
\- Copy and pasting the url that will appear in your Terminal after successfully pushing your code to the remote Heroku branch
\- Navigating to your project on [the Heroku website](http://heroku.com/) and there will be a button that says 'View App' - click it and it will take you there
* If you see the 'Hello from Express!' (or whatever test message you used) then that means your server is running correctly
* * *
- Now for our MongoDB connection to work we must define an environment variable for Heroku to store our MongoDB connection string.
> We want to keep our connection string secret as we don't want anybody able to connect to our database and change, delete, add things etc.
> In our local build the connection string, is stored in our `.env` file inside of `/server`. I named by connection string `MONGODB_URI`. You can do this next step on the Heroku CLI or on the Heroku website.
> I recommend the website it's more straightforward and you don't have to worry about escaping special characters. I'm going to describe the next steps going through the website.
* Navigate to your dashboard on Heroku
* Click on your project name that will be present from our previous steps where we created our remote branch and pushed the code
* Navigate to the Settings tab near the top
* The second section is 'Config Vars' - click on 'Reveal Config Vars'
* You'll see two inputs:
* One is the name of your environment variable - name this **whatever you have it named in your local build**. For me that is `MONGODB_URI`.
* Second is the value - paste your whole connection string here that you should copy from your `.env` file directly to avoid any typos.
* Then just click 'Add' and our MongoDB connection environment variable is set. 
* * *
Our `server` code is officially deployed and configured correctly. Now onto the `client` code with Netlify.
* * *
-
Next is to deploy the front-end React code in
/client
to Netlify. Login to Netlify if you haven't already, the default tab should beSites
or just navigate to that tab. -
LASTLY after we deploy our front-end React code we must ensure any requests we're sending from the client-side is changed to use our Heroku URL now instead of localhost.
* In my structure the requests were being made from `client/api/index.js` so navigate to that file and _any request that contains_ `http://localhost:5000` must be replaced by your Heroku URL.
// Before
const res = await fetch('http://localhost:5000/scores/'
// After
const res = await fetch('https://my-project.herokuapp.com/scores/')
// 'my-project' in the URL will either be the auto-generated
// name from Netlify or if you changed the name it will
// be the name you gave it
Ensure that you push these changes up to GitHub. Netlify will trigger a redeploy when they detect changes to your master
branch. So for this to work you must make those changes apparent to Netlify essentially.
Now any request, instead of going to your local server you've ran while developing, will go to the hosted server you just deployed to Heroku with all of your Express Routes still intact and functioning properly.
It's undoubtedly a long process to get this done. But the important thing is that it can be done!
It's awesome to see the things we build on our local machine in development. After this though you can send a live link to your friends, colleagues, etc. for them to check it out!
This is awesome to have a live example as well for anyone looking for a job. Being able to not only have a link on your Resume or Portfolio site to the code on GitHub but for recruiters, managers, whoever to be able to see the project in action is huge!
... how you like to deploy! Link me an article, give me a brief overview, or whatever you want. I'd love to hear of the other ways devs get their projects out there for the world to see and enjoy!