Deploying Red Hat JBoss Fuse Using ACS and Kubernetes

Red Hat JBoss Fuse has been the de facto standard for building Java web/RESTful services for over a decade. But how do you run it effectively in today’s cloud-centric world? As you’ll see, an infrastructure-as-a-code and scalable/fault-tolerant approach are both critical for a successful deployment.

In this tutorial, we’ll show you how to:

1. Build an environment in a Kubernetes (K8s) cluster in Azure.

2. Package your Red Hat JBoss services into a Docker Container.

3. Run your services in a scalable, highly-available cluster.

Building an Environment on a Kubernetes Cluster in Azure

To start, you’ll need an operational Kubernetes cluster.

After installation, run the Kubernetes Dashboard (kubectl proxy) and ensure the Kubernetes Dashboard UI (http://127.0.0.1:8001/ui) is working:

Packaging Your Red Hat JBoss Services Into a Docker Container

The typical Red Hat JBoss deployment process requires you to install Red Hat JBoss, configure Red Hat/Karaf features, and deploy your services (i.e., the developed *.jar files). You can automate the installation with the Docker file and get a delivery unit that is ready for testing and deployment to production.

Docker file:

Set up your Docker Image Registry (or use DockerHub), and configure your Docker to access the Registry.

We use the commands “bin/client log:display” in the Docker file to help ensure that the Red Hat reconfigurations and deployments were successful.

After this, the typical developer’s flow will be to build a Docker Container Image, tag the Image with a version, and push the Image to the Docker Registry:

Running Your Services in a Scalable, Highly Available Cluster

You have now successfully configured Kubernetes on Azure Container Service, and you have a Docker Image in a Docker Registry. Next, you’re ready to proceed with Kubernetes!

Basically, you will need to create one Deployment (for Red Hat nodes) and one Service (for a load balancer and publicly accessible IP) in Kubernetes.

To create the Deployment, go to the dashboard, and select “Deployment” in the left menu.

Click “+Create” in the upper right, and select the “Upload a YAML or JSON file” option.

Here is the Kubernetes deployment definition:

Select Rh-deployment.yaml, and click “Upload”. Next, select “Pods” in the left menu. Wait until “rhesb-….” is ready. This should take approximately 5 minutes because the image is 2GB.

Then go to “Services”, click “+Create”, and select the Kubernetes service definition(rh-service.yaml). Ensure the “selector” attribute in the Service definition matches the “metadata” name of the Deployment definition.

Here is the Kubernetes service definition:

Go to “Services” and wait until the new service displays an IP address. This will take a few minutes as the load balancer and rules are created.

You are now ready to test your service using SoapUI, or a similar tool:

While you can go into production with this Docker Image and a couple of Kubernetes YAML files, you should also:

• Choose an approach for managing environment-specific properties (for example, URLs and ports of services).

• Set up log shipping using Stash or the Azure Monitoring Agent.

• Add readinessProbe and livenessProbe (for each service) to Kubernetes Service definition to ensure you are not the owner of a cluster of all-dead nodes.

Share your thoughts and questions in the comments section below.