Moving code from development to production doesn't have to be as error-prone and time-consuming as it often is. By using Buddy, a continuous integration and delivery tool that doubles up as a powerful automation platform, you can automate significant portions of your development workflow, including all your builds, tests, and deployments.
Unlike many other CI/CD tools, Buddy has a pleasant and intuitive user interface with a gentle learning curve. It also offers a large number of well-tested actions that help you perform common tasks such as compiling sources and transferring files.
In this tutorial, I'll show you how you can use Buddy to build, test, and deploy a Node.js app.
Prerequisites
To be able to follow along, you must have the following installed on your development server:
1. Setting Up a Node.js App
Before you dive into Buddy, of course, you'll need a web app you can build and deploy. If you have one already, feel free to skip to the next step.
If you don't have a Node.js app you can experiment with, you can create one quickly using a starter template. Using the popular Hackathon starter template is a good idea because it has all the characteristics of a typical Node.js app.
Fork the template on GitHub and use git to download the fork to your development environment.
git clone https://github.com/hathi11/hackathon-starter.git
It's worth noting that Buddy is used with a Git repository. It supports repositories hosted on GitHub, BitBucket, and other such popular Git hosts. Buddy also has a built-in Git hosting solution or you can just as easily use Buddy with your own private Git servers.
Once the clone's complete, use npm to install all the dependencies of the web app.
cd hackathon-starter/ npm install
At this point, you can run the app locally and explore it using your browser.
node app.js
Here's what the web app looks like:
2. Creating a Buddy Project
If you don't have a Buddy account already, now is a good time to create one. Buddy offers two premium tiers and a free tier, all of which are cloud based. The free tier, which gives you 1 GB of RAM and 2 virtual CPUs, will suffice for now.
Once you're logged in to your Buddy account, press the Create new project button to get started.
When prompted to select a Git hosting provider, choose GitHub and give Buddy access to your GitHub repositories.
You should now be able to see all your GitHub repositories on Buddy. Click on the hackathon-starter repository to start creating automations for it.
Note that Buddy automatically recognizes our Node.js app as an Express application. It's because our starter template uses the Express web app framework.
3. Creating a Pipeline
On Buddy, a pipeline is what allows you to orchestrate and run all your tasks. Whenever you need to automate something with Buddy, you either create a new pipeline for it or add it to an existing pipeline.
Click on the Add a new pipeline button to start creating your first pipeline. In the form shown next, give a name to the pipeline and choose On push as the trigger mode. As you may have guessed, choosing this mode means that the pipeline is executed as soon as you push your commits to GitHub.
The next step is to add actions to your pipeline. To help you get started, Buddy intelligently generates a list of actions that are most relevant to your project.
For now, choose the Node.js action, which loads a Docker container that has Node.js installed on it. We'll be using this action to build our web app and run all its tests. So, on the next screen, go ahead and type in the following commands:
npm install npm test
4. Attaching a Service
Our web app uses MongoDB as its database. If it fails to establish a connection to a MongoDB server on startup, it will exit with an error. Therefore, our Docker container on Buddy must have access to a MongoDB server.
Buddy allows you to easily attach a wide variety of databases and other services to its Docker containers. To attach a MongoDB server, switch to the Services tab and select MongoDB. In the form shown next, you'll be able to specify details such as the hostname, port, and MongoDB version you prefer.
Make a note of the details you enter and press the Save this action button.
Next, you must configure the web app to use the URI of Buddy's MongoDB server. To do so, you can either change the value of the MONGODB_URI field in the .env.example file, or you can use an environment variable on Buddy. For now, let's go ahead with the latter option.
So switch to the Variables tab and press the Add a new variable button. In the dialog that pops up, set the Key field to MONGODB_URI and the Value field to a valid MongoDB connection string that's based on the hostname you chose earlier. Then press the Create variable button.
The official documentation has a lot more information about using environment variables in a Buddy pipeline.
5. Running the Pipeline
Our pipeline is already runnable, even though it has only one action. To run it, press the Run pipeline button.
You will now be taken to a screen where you can monitor the progress of the pipeline in real time. Furthermore, you can press any of the Logs buttons (there's one for each action in the pipeline) to take a closer look at the actual output of the commands that are being executed.
You can, of course, also run the pipeline by pushing a commit to your GitHub repository. I suggest you make a few changes to the web app, such as changing its header by modifying the views/partials/header.pug file, and then run the following commands:
git add . git commit -m "changed the header" git push origin master
When the last command has finished, you should be able to see a new execution of the pipeline start automatically.
6. Moving Files
When a build is successful and all the tests have passed, you'd usually want to move your code to production. Buddy has predefined actions that help you securely transfer files to several popular hosting solutions, such as the Google Cloud Platform, DigitalOcean, and Amazon Web Services. Furthermore, if you prefer using your own private server that runs SFTP or FTP, Buddy can directly use those protocols too.
In this tutorial, we'll be using a Google Compute Engine instance, which is nothing but a virtual machine hosted on Google's cloud, as our production server. So switch to the Actions tab of the pipeline and press the + button shown below the Build and test action to add a new action.
On the next screen, scroll down to the Google Cloud Platform section and select the Compute Engine option.
In the form that pops up, you must specify the IP address of your VM. Additionally, to allow Buddy to connect to the VM, you must provide a username and choose an authentication mode.
The easiest authentication mode in my opinion is Buddy's SSH key. When you choose this mode, Buddy will display an RSA public key that you can simply add to your VM's list of authorized keys.
To make sure that the credentials you entered are valid, you can now press the Test action button. If there are no errors, you should see a test log that looks like this:
Next, choose GitHub repository as the source of the files and use the Remote path field to specify the destination directory on the Google Cloud VM. The Browse button lets you browse through the filesystem of the VM and select the right directory.
Finally, press the Add this action button.
7. Using SSH
Once you've copied the code to your production server, you must again build and install all its dependencies there. You must also restart the web app for the code changes to take effect. To perform such varied tasks, you'll need a shell. The SSH action gives you one, so add it as the last action of your pipeline.
In the form that pops up, you must again specify your VM's IP address and login credentials. Then, you can type in the commands you want to run. Here's a quick way to install the dependencies and restart the Node.js server:
pkill -HUP node #stop node server cd my_project npm install #install dependencies export MONGODB_URI= nohup node app.js > /dev/null 2>&1 & #start node server
As shown in the Bash code above, you must reset the MONGODB_URI environment variable. This is to make sure that your production server connects to its own MongoDB instance, instead of Buddy's MongoDB service.
Press the Add this action button again to update the pipeline.
At this point, the pipeline has three actions that run sequentially. It should look like this:
Press the Run pipeline button to start it. If there are no errors, it should take Buddy only a minute or two to build, test, and deploy your Node.js web app to your Google Cloud VM.
Conclusion
Being able to instantly publish new features, bug fixes, and enhancements to your web apps gives you a definite edge over your competition. In this tutorial, you learned how to use Buddy's pipelines, predefined actions, and attachable services to automate and speed up common tasks such as building, testing, and deploying Node.js applications.
There's a lot more the Buddy platform can do. To learn more about it, do refer to its extensive documentation.
No comments:
Post a Comment