In this sprint, you will learn about basic deployment and various build tools. The tools and techniques you gain experience with here will allow you to kick off your group projects with a bang.
Deployment, at times, can be really painful. In particular, deployment bugs often take longer cycles to investigate than local code, which can feel very frustrating, and, deployment bugs may often feel like they are just nuances to someone else's tools that they did not take the time to document well. This is in fact often true, but it comes with the territory, and learning to enjoy the perhaps slower-seeming process will help you develop a set of tools that gives you tremendous power as an engineer.
In the spirit of deployment you will find many of the instructions given at a rather high level. It will be up to you and your pair to orient and figure out the details of how to do what needs doing.
Learn how to deploy on Heroku Learn how to use Liberator Learn how to deploy on DigitalOcean Feel some pain around deployment bugs and the pace of deployment Gain some rudimentary exposure to how build tools (like Grunt) can make your life easier Gain some exposure to MongoDB and Mongoose
Heroku is widely popular for good reason. Getting started with it is incredibly easy, and the documentation is just fantastic. Many developers and startups will begin with Heroku, allowing them to focus on rapid development, potentially moving on to more sophisticated deployment options only when the need arises.
Learning to get started with Heroku is not such an interesting challenge, but you should have it as a tool in your toolbelt. Expect to spend about an hour and no more working with Heroku before moving onto deploying Shortly on DigitalOcean.
Complete Getting Started with Node.js on Heroku
We're giving you a canonical version of the shortly-express repo to start with. Before diving in, do a code review, out loud with your pair. Take a few minutes with your pair to compare this canonical repo to your work from the last sprint. How was your app architected differently? Could your code have been DRYer? Are there any functional differences between your apps and this one? There are many many ways to write code, and you can often learn as much from reading code as you can from writing it.
You will find all tests in this repo are in a pending state. They are all written for MongoDB. Before you begin the MongoDB refactor later in the sprint, remove 'x' from each describe block in the tests.
This repo uses supertest (see the require block in test/ServerSpec.js), a powerful library for testing HTTP. Use it to write at least 2 additional tests for the application
Fork the shortly-deploy repo Use Liberator on your already existing fork of shortly-deploy. It will result in you having a new fork that is fit for deployment to the outside world Clone down your liberated repo npm install, node server.js and make sure you can access the shortly code on localhost:4568
Create an account with some appropriate billing information Set up a billing alert so that you receive an email should your monthy usage exceed some amount you decide. Feel free to make this amount as small as you feel necessary
Most deployment services, DigitalOcean included, allow you to pay them for giving you access to a Linux server that has an IP address exposed to the world wide web. Generally, you house and run your application code inside the Linux server you are paying to use so that others can access it. Often times, once you have access to one of these deployment servers, you have to install, in addition to your application code, programs and code that your application code needs, like, for example, NodeJS, git, or MongoDB.
DigitalOcean calls the Linux servers they offer "droplets," and you can create many different kinds of droplets that already contain programs and code that might be useful for your application.
Create a new DigitalOcean droplet, using the MEAN on 14.04 image in the "One-click Apps" section of the images. Choose the $10 droplet, in any datacenter, and give it any hostname you like. Do not select any additional options. Do be sure to create an SSH key for it. It will take about a minute to create the droplet, at which point in time you will be given the external facing IP address of the droplet.
SSH or Secure Shell, is a protocol that allows you to securely log into remote servers from the terminal. Typically, when using SSH, you will need to be explicit about which user you are logging in as. In order to use SSH, from inside the terminal on your local machine, you issue the following command:
1│ssh @ SSH into the droplet you created as user root, answering yes to the question you are probably asked. You will not be able to do this if you did not successfully create an SSH key in the step above Once inside your droplet, clone down the liberated shortly-deploy repo somewhere inside it. Do npm install Now run node server.js and hit the correct port and IP address from inside your browser to see your liberated shortly-deploy app deployed to the world. Address any bugs
There are many services out there to help make it easy on you, the developer, to continually deploy newer versions of your code. Remember how easy git push heroku master was? You are now going to roll your own version of this for your DigitalOcean deployment by way of a somewhat deep dive into git. The following task will require that you balance getting what you need to know out of the resource vs. going down a rabbit hole.
Read How to Set Up Automatic Deployment with git with a VPS, up to the section titled Beta Having SSHed into your droplet, create some directory (not inside your cloned down code) to create a "bare" git repo, by issuing the git init --bare command from inside it Taking a lead from the article you just read, create and edit a post-receive file inside the hooks directory of the "bare" repo that sets the --work-tree option to the directory where you have cloned down your codebase, and the --git-dir option to the location of the "bare" repo that you created. Be sure to chmod +x the post-receive file you just created Again taking a lead from the article above, on your local machine, create a live remote inside your shortly-deploy repository that points to the "bare" repo in your droplet, as the root user Make some change to your local shortly-deploy repo on the master branch, commit the changes, and then git push live master. If you are successful the command line will indicate it, and, you should be able to observe the changes when SSHed into to your droplet
Building and deploying an app involves a number of important tasks that need to be performed in a certain order. When you're trying to rapidly prototype your app, this can become repetitive and is prone to error. Grunt is one of many super useful tools that can automate a wide variety of tasks for you.
Use Grunt to create tasks that do the following:
Start your server locally using nodemon Use the live remote to push your local code up to your deployed droplet. Assuming you've already started your deployed code with nodemon you should be able to git push live master and watch your deployed code base update to the world Concatenate files before deployment Uglify your code before deployment. Don't forget to update your views to point to the minified versions of your css and js asset files in the public/dist folder. The folder public/dist is already .gitignored for you, but make sure that you aren't committing your compiled scripts and/or css to your Github repo Run eslint before deployment. If eslint fails, the build process should exit Run your Mocha tests before deployment. If any tests fail, the build process should exit Watch your source code for changes in order to rerun any of the Grunt tasks you created as is appropriate When you're done, feel free to take up to 20 minutes to combine many of your tasks to run with a single command.
Create a grunt deploy task that utilizes tasks already created in order to build and run your app locally Add a prod option such that when you run grunt deploy --prod you will prepare your code base for production and push it up to the production droplet
In the previous sprint, our shortened links were stored using sqlite, a server-less database engine. Sqlite is great for many things, but is not the best choice in most cases for well-trafficked production sites for various reasons.
MongoDB, on the other hand, is one of the most widely used databases in production environments and offers the most popular alternative to using an RDBMS.
Remove the 'x' from each describe block in the tests Refactor the app to use MongoDB and Mongoose, making the tests pass and checking your application locally. Be sure to use promises. If the built-in promises are too limiting for you, plug in your preferred library. If you need to install MongoDB on your laptop, follow the installation instructions. Assuming you used the MEAN droplet, your deployed DigitalOcean server already has a mongod instance running on it. Push your refactored code up to production and profit
Our advanced content is intended to throw you in over your head, requiring you to solve problems with very little support or oversight, much like you would as a mid or senior level engineer.
Several routes in the server use nested callbacks. Refactor them all to use promises. Bluebird, a popular and performant promise library, is already included
Deploy your site to another service of your choosing: Azure AWS is massive, and you should give it a try if ready to get in a little over your head. The best place to get started learning about AWS is by taking taking the AWS quickLab on EC2
Install the Docker Toolbox on your local machine Use docker-machine (installed above) to create "machines" on DigitalOcean Use a Dockerfile, with the node image as a base, to create a Docker image for your shortly web server Use the mongo docker image to launch a mongo container, networking your shortly web server container to it Compose the two containers and launch them both on DigitalOcean Use your newfound Docker prowess to roll your own CD
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent