Building server images with Docker using Google Cloud Build. Infrastructure as code is a blanket term that covers many different technologies and methods for writing configuration files, and then deploying them as servers on different platforms. While we can use a traditional virtual server in an IaC deployment, we might be tempted to lean on old habits of treating it like a "pet" server and not as "livestock." We are better off using a container-based solution that will force us to treat the server as an immutable entity.

Containers are like lightweight virtual machines, but they are stateless with no persistent storage and are rebuilt from a configuration file every time changes are made. So you can see containers are basically synonymous with infrastructure as code. Docker is one of the most widely used tools for building containers. We can build a Docker image that will run our custom server by defining just a single configuration file. This file is called rather unimaginatively a Docker file.

Let's take a look at an example Docker file, make some changes to it and see how we can use Google Cloud Build to create our customized server image and store it on GCP for us. For this course, we are going to create infrastructure as code for a WordPress deployment.

So we'll start with the official base WordPress Docker image and build our Docker image on top of that. We can see in the WordPress Docker file, all the commands it already runs to prepare the WordPress server for us and in the very first line we can see how this image itself is built from the official PHP Docker image.

We can take a look at the PHP Docker file then, and see that this version includes the Apache web server for us and that this image is built from a base Debian operating system container image. Knowing that our server image combines all of these Dockerfiles, we are able to add or change any configuration settings for the Debian, PHP or WordPress containers in our own Docker file to customize our server image.

Very often, we may need to add a PHP extension required by a particular WordPress plugin or change a PHP configuration setting to improve performance with WordPress. We can include our custom php.ini configuration with our Docker file and write this command in our Docker file to copy our configuration into our build.

The PHP Docker repository comes with some helper scripts for adding and removing PHP extensions. But to make this even easier, we are adding the Docker PHP extension installer repository. This way we can enable xdebug and a couple other extensions in PHP with just a single command and it will handle any dependencies for us.

Remember that containers are immutable with no persistent storage but we are running a WordPress server and we'll definitely want to be able to add content and save changes to our website. We can do this by creating a SQL database and a persistent storage disc outside of our container and then mapping them to our server image. We'll learn how to tie all of these elements together later with Google Deployment Manager.

Now that we've created our custom WordPress Docker file, we can build our new WordPress server image and store it in Google Artifact Registry with a simple Docker command. We could instead use the cloudbuild.yml file from our repository that we used with our Cloud Build Trigger earlier like this.

Any command-line arguments that we would pass to the G Cloud Build command can be instead stored in a cloudbuild.yml config file like this which allows us to add variables that make our configurations more easily reusable like we are doing here with the project ID. It's important to remember number that there are costs incurred when building images with Google Cloud Build like we demonstrated here.

You could instead install Docker on your development machine, use the Docker build command locally then upload the same resulting image to the artifact registry in order to avoid the expense of cloud build. Depending on your situation and workflow, there may be good reasons to use cloud build though. If you are unable to install Docker locally, if your development machine is an underpowered code typewriter and can't handle building images, or if you have a lousy internet connection and don't want to wait an hour for your images to upload between updates, Cloud Build could very well be worth the cost.