Habitat, Google Container Registry (GCR) and Google Kubernetes Engine (GKE)

This short post will explore how we can use Habitat and Google Container Registry to deploy a managed container to Google Kubernetes Engine (GKE) with Habitat Operator already installed.

The example we will be following is here. Checking this out reveals the following structure:

$ tree hab-gcr-demo
├── config
│   ├── index.html
│   └── nginx.conf
├── default.toml
├── deploy-hab-gcr-demo.yml
├── hooks
│   ├── init
│   └── run

From a Habitat perspective, the key lines of the are as follows:


do_build() {
  return 0

do_install() {
  return 0

This means we depend on the core/nginx maintained package and don’t require any special build or installation setup. This package will export a port which is important for us to see the index.html page when deployed at a later stage. The configuration directory also includes a simple nginx.conf file.

Build the Habitat package and create the container image

With Habitat installed locally, let’s first trigger a build:

$ cd hab-gcr-demo/
$ hab pkg build .
   hab-studio: Creating Studio at /hab/studios/src (default)

This will create a hart file in the results folder according to the origin and package name e.g.

$ ls results/
habskp-hab-gcr-demo-0.1.0-20180711153450-x86_64-linux.hart	logs

From the Studio or otherwise, now execute the following to create a container image. This is what we will upload to Google Container Registry.

$ hab pkg export kubernetes results/habskp-hab-gcr-demo-0.1.0-20180711153450-x86_64-linux.hart

This creates a Docker image with tags e.g.

$ docker images
REPOSITORY                                  TAG                     IMAGE ID            CREATED             SIZE
habskp/hab-gcr-demo                         0.1.0                   ed2580b52a3a        53 seconds ago      271MB
habskp/hab-gcr-demo                         0.1.0-20180711153450    ed2580b52a3a        53 seconds ago      271MB
habskp/hab-gcr-demo                         latest                  ed2580b52a3a        53 seconds ago      271MB

Push the container image to GCR

The following command ensures we are authorized to push/pull images to Container Registry:

$ gcloud auth configure-docker

Now we run the following two commands to tag and push our image, note that the repository URI conforms to [HOSTNAME]/[PROJECT-ID]/[IMAGE]:[TAG], more details here:

$ docker tag habskp/hab-gcr-demo:latest
$ docker push
The push refers to repository []
567f6430b59f: Pushed 
latest: digest: sha256:a9db7cc0cf186311eddde7d5f796b998b6aad5521d435ffa63ee572d0fb1d73a size: 529

Deploy your application to GKE

With our image in the Container Registry we can deploy the application to Google Kubernetes Engine using the provided manifest:

$ cat deploy-hab-gcr-demo.yml 
kind: Habitat
  name: hab-gcr-demo
customVersion: v1beta2
    count: 1
      name: hab-gcr-demo
      topology: standalone
apiVersion: v1
kind: Service
  name: hab-gcr-demo-lb
  type: LoadBalancer
  - name: web
    port: 80
    targetPort: 8080
    habitat-name: hab-gcr-demo

This also creates a Kubernetes LoadBalancer to expose port 8080 from the container running nginx to a public IP address on port 80. This manifest can now be deployed to GKE:

$ kubectl apply -f deploy-hab-gcr-demo.yml 
habitat "hab-gcr-demo" created
service "hab-gcr-demo-lb" created
$ kubectl get services hab-gcr-demo-lb
NAME              TYPE           CLUSTER-IP     EXTERNAL-IP      PORT(S)        AGE
hab-gcr-demo-lb   LoadBalancer   10.451.200.92   80:31696/TCP   1m

Navigating to the page we see the expected index.html:

Successful Index Page


In this walkthrough you saw how to push a Habitat managed application to GCR and how to deploy it to GKE. Watch this space, Habitat Builder integration for Google Container Registry is coming soon. Thanks for reading!

Got questions?

Read more:

Posted in:

Stuart Paterson

Stuart Paterson is a Principal Engineer at Chef Software. He is based in Belfast and enjoys solving technical challenges in close collaboration with Chef’s partners. Before joining Chef, Stuart worked in a variety of roles across multiple verticals in Financial Services. His background before that was in particle physics.