6.0 KiB

Raw Blame History

Example Application

TL;DR

$ kubectl create -f https://raw.githubusercontent.com/bitnami/bitnami-docker-ruby/master/example/kubernetes.yml

Introduction

This example demostrates the use of the bitnami/ruby image to create a production build of your ruby application.

For demonstration purposes we'll bootstrap a Rails application, build a image with the tag bitnami/ruby-example and deploy it on a Kubernetes cluster.

Generate the application

The example application is a Rails application bootstrapped using the rails new command.

$ rails new example --skip-active-record --skip-bundle

Build and Test

To build a production Docker image of our application we'll use the bitnami/ruby:2.4-prod image, which is a production build of the Bitnami Ruby Image optimized for size.

FROM bitnami/ruby:2.4 as builder
ENV RAILS_ENV="production"
COPY . /app
WORKDIR /app
RUN bundle install --no-deployment
RUN bundle install --deployment
RUN bin/rails generate controller Welcome index
RUN bin/bundle exec rake assets:precompile


FROM bitnami/ruby:2.4-prod
ENV RAILS_ENV="production" \
    SECRET_KEY_BASE="your_production_key" \
    RAILS_SERVE_STATIC_FILES="yes"
RUN install_packages libssl1.0.0
COPY --from=builder /app /app
WORKDIR /app
EXPOSE 3000
CMD ["bin/rails", "server"]

The Dockerfile consists of two build stages. The first stage uses the development image, bitnami/ruby:2.4, to copy the application source, install the required gems using bundle install, generate a dummy controller and precompile the assets. The RAILS_ENV environment variable is defined so that bundle install only installs the application gems that are required in production executions and also for the rails server to start in production mode.

The second stage uses the production image, bitnami/ruby:2.4-prod, and copies over the application source and the installed gems from the previous stage. This creates a minimal Docker image that only consists of the application source, gems and the ruby runtime.

To build the Docker image, execute the command:

$ docker build -t bitnami/ruby-example:0.0.1 example/

Since the bitnami/ruby:2.4-prod image is optimized for production deployments it does not include any packages that would bloat the image.

$ docker image ls
REPOSITORY               TAG                 IMAGE ID            CREATED             SIZE
bitnami/ruby-example     0.0.1               847d58b5bc8a        4 minutes ago       203MB

You can now launch and test the image locally. You will need to access to http://YOUR_IP:3000/welcome/index

$ docker run -it --rm -p 3000:3000 bitnami/ruby-example:0.0.1

=> Booting Puma
=> Rails 5.1.4 application starting in production
=> Run `rails server -h` for more startup options
Puma starting in single mode...
* Version 3.10.0 (ruby 2.4.2-p198), codename: Russell's Teapot
* Min threads: 5, max threads: 5
* Environment: production
* Listening on tcp://0.0.0.0:3000
Use Ctrl-C to stop

Finally, push the image to the Docker registry

$ docker push bitnami/ruby-example:0.0.1

Deployment

The kubernetes.yml file from the example/ folder can be used to deploy our bitnami/ruby-example:0.0.1 image to a Kubernetes cluster.

Simply download the Kubernetes manifest and create the Kubernetes resources described in the manifest using the command:

$ kubectl create -f kubernetes.yml
ingress "example-ingress" created
service "example-svc" created
configmap "example-configmap" created
persistentvolumeclaim "example-data-pvc" created
deployment "example-deployment" created

From the output of the above command you will notice that we create the following resources:

Note

Our example application is stateless and does not store any data or does not require any user configurations. As such we do not need to create the ConfigMap or PersistentVolumeClaim resources. Our kubernetes.yml creates these resources strictly to demostrate how they are defined in the manifest.

Accessing the application

Typically in production you would access the application via a Ingress controller. Our kubernetes.yml already defines a Ingress resource. Please refer to the Ingress documentation to learn how to deploy an ingress controller in your cluster.

Hint

https://kubeapps.com/charts/stable/nginx-ingress

The following are alternate ways of accessing the application, typically used during application development and testing.

Since the service example-svc is defined to be of type NodePort, we can set up port forwarding to access our web application like so:

$ kubectl port-forward $(kubectl get pods -l app=example -o jsonpath="{ .items[0].metadata.name }") 3000:3000

The command forwards the local port 3000 to port 3000 of the Pod container. You can access the application by visiting the http://localhost:3000/welcome/index.

Note:

If you are using minikube, you can access the application by simply executing the following command:
$ minikube service example-svc

Health Checks

The kubernetes.yml manifest defines default probes to check the health of the application. For our application we are simply probing if the application is responsive to queries on the root resource.

You application can define a route, such as the commonly used /healthz, that reports the application status and use that route in the health probes.

6.0 KiB Raw Blame History