Something to work with

In order to be able to do much in GCP we need to have Docker images that represent our application that we want work with. To make this showcase a bit more interesting than a classic 'Hello World App' pingpong-service has been built for you. It is special designed to display a set of capabilities that comes with GCP and show how easy it is to make use of them from your application.

There are many ways to produce Docker images and many pipelines to do the job but in this example we will use the Cloud Build in combination with the Container Registry

Cloud Build

Cloud Build is a function in GCP that allows you to specify declarative step-by-step build pipelines. Each build step in your pipeline is performed by a container doing some operation on your workspace. The workspace is shared across the steps but not between different builds. There are predefined build step containers, like for Maven, Git, Docker and so on, but you can also create your own custom made build step containers and use them in Cloud Build. For us, in this example, the predefined ones will be more than enough!

Container Registry

A bit sloppy you could say that Container Registry is GCP:s version of Docker Hub. It's not all true of course but we will only use it to store the Docker images that we produce during build. The images will be private to us and directly accessible within our project which is nice! We could push our images to Docker Hub instead if we wanted to but then we'd have to provide Docker Hub credentials etc.

Producing the Docker images

The repo that you cloned in the previous step, pingpong-service, comes with a cloudbuild.yaml-file in the root folder. It looks more or less like this:

steps:
- name: 'gcr.io/cloud-builders/mvn'
  args: ['install']
- name: 'gcr.io/cloud-builders/docker'
  args: ['build', '--tag=gcr.io/$PROJECT_ID/pingpong-service-simple', 'pingpong-service-simple']
- name: 'gcr.io/cloud-builders/docker'
  args: ['build', '--tag=gcr.io/$PROJECT_ID/pingpong-service-postgres', 'pingpong-service-postgres']
- name: 'gcr.io/cloud-builders/docker'
  args: ['build', '--tag=gcr.io/$PROJECT_ID/pingpong-service-postgres-redis', 'pingpong-service-postgres-redis']
images: ['gcr.io/$PROJECT_ID/pingpong-service-simple', 
         'gcr.io/$PROJECT_ID/pingpong-service-postgres', 
         'gcr.io/$PROJECT_ID/pingpong-service-postgres-redis']

As you can see it contains several steps.

The first is a maven build step that will build all the different flavours of pingpong-service that we will use (refer to the README.md of pingpong-service for more info about those). It will use the maven install command which will build code and produce JAR-files, in this case Spring Boot fat-JAR:s.

The following steps are 3 different docker build commands that takes the 3 produced fat-JAR:s and put them in a Docker image each.

The last line 'images' pushes the 3 produced Docker images to Container Registry making them available for use.

Now, this is not your Github project so you can't attach the repo to Cloud Build and use it like it was your own, but instead you can just run the cloudbuild.yaml file locally using the Google Cloud SDK. This will upload the source files to Cloud Build of your project and execute the actual build there.

If you really want to try Github integrations with Cloud Build you can of course fork this project to your Github account and make integrated use of it in Google Cloud.

Go to the root folder of your pingpong-service project if you are not already there and type:

gcloud builds submit --config=cloudbuild.yaml

If you have your Google Cloud SDK properly installed this should start the build and it will take a minute or two to perform. You can monitor progress in your terminal window or in the Google Cloud Build logs. Once the build is complete and successful go to Container Registry and Images and you should see something similar to the following:

Ok?

Well done, we can now proceed to the next step.