Deploying a MEAN Stack Onto Google Compute Engine

Gone are the days where a product team used to spend a considerable amount of time to build a basic web application.  Say Hi to MEAN! If you are familiar with LAMP/WAMP stacks, you could consider MEAN as a complete stack based on JavaScript. In fact, MEAN represents MongoDB,  Express, AngularJS, and NodeJS:

  • Mongo DB – The Leading NoSQL database
  • Express JS – The Web Application Framework
  • Angular JS  – The HTML enhanced Web Apps
  • Node JS  – a platform to build scalable network applications

With MEAN, you could have your web application running in less than 10 minutes.
Deploying a MEAN based application on AWS or Azure is quite straightforward as there are ready-made templates from 3rd party vendors like Bitnami. Nevertheless, it’s not that easy on Google Compute Engine, given that there are no built-in images for that platform.

In this post, I will describe the rationale in choosing MEAN stack over others, and 10 Steps to deploy a MEAN stack on Google Compute Engine
We have had rich experience in creating many enterprise application using J2EE, Python, .Net and other technologies. You may wonder why on earth we chose MEAN! Let me list down the reasons:

  • JavaScript is often considered to be front end technology. Actually, everything can be written in JavaScript and there are more and more examples of effective server-side usage of Javascript.
  • We want to use an open source language, with strong community support and with a quick development cycle. We had many options, like Python, Ruby and Angular + Node. We had had to learn different languages for UI and Backend in case of Python and Ruby. But Javascript is used in both Angular and Node.
  • AngularJS supports a native MVC framework and promotes Test Driven Development with the Karma/Mocha test framework. AngularJS UI renders on all kinds of browsers supporting HTML5 supported.
  • ExpressJS is a good web framework on top of NodeJS. It provides us hooks to easily develop a scalable application server.

So, all this considered, here are the steps to deploy a MEAN stack on GCP
Step 1: Sign up for Google Compute Engine, or sign in if you have an account already

Step 2: Create a project or Go into your existing Project

Step 3: Go to Google Compute Engine -> VM Instances

Step 4: Create a CentOS based VM Instance

Step 5: Once the VM is up, SSH into the VM Instance
gcutil --service_version="v1" --project="##your_project" ssh --zone="asia-east1-a" "my-mean-stack"

Step 6: Download NodeJS and Place it on your path:
wget http://nodejs.org/dist/v0.10.28/node-v0.10.28-linux-x64.tar.gz
tar xvfz node-v0.10.28-linux-x64.tar.gz
mv node-v0.10.28-linux-x64 /usr/share/

Step 7: Install MongoDB and git:
sudo yum install mongodb-org git

Step 8: Deploy meanio module
$sudo npm install -g meanio
$mean init my-mean-app
Cloning branch: master into destination folder: my-mean-app
Initialized empty Git repository in /home/user/my-mean-app/.git/
....
install dependencies:
$ cd my-mean-app && npm install

Step 9: Start your server
$ grunt
Running "clean:0" (clean) task

Running "jshint:all" (jshint) task
>> 41 files lint free.

Running "csslint:src" (csslint) task
>> 3 files lint free.

Running "concurrent:tasks" (concurrent) task
Running "watch" task
Waiting...
Running "nodemon:dev" (nodemon) task
[nodemon] v1.0.20
[nodemon] to restart at any time, enter `rs`
[nodemon] watching: *.*
[nodemon] starting `node --debug server.js`
debugger listening on port 5858
js-bson: Failed to load c++ bson extension, using pure JS version
Mean app started on port 3000 (development)
Failed to load c++ bson extension, using pure JS version

Step 10: Open up port 3000 in your Google Compute Network Firewall

TThat’sit! Launch your browser and point it to http://your_external_ip:3000/. Your web app is up and running!

Cloud Academy