This means Fabric8 reuses the native container orchestration that’s baked into:
- Google Container Engine, OpenShift, Azure
- RHEL Atomic (the host linux distro optimised for running Docker containers)
- Jube which is a pure Java implementation of Kubernetes which works on any platform that has a JVM (by emulating Docker via running processes using image zips; which if you squint are similar to docker images just without the linux operating system image or isolation & resource limiting)
- hopefully someone figures out how to support the Kubernetes APIs natively on EC2 Container Service so folks can use that too; otherwise its pretty easy to run Kubernetes on EC2 directly
This lets you make reusable enterprise applications which work on all Kubernetes platforms which can easily be shared such as by dragging and dropping an App Zip onto your Library on the Fabric8 Console.
To get started you need to pick which Kubernetes platform you want to use based on your needs:
- If you have Docker installed then try Getting Started with Fabric8 and Docker which uses OpenShift V3 in a docker image as the Kubernetes implementation.
- If you want to simulate a multi-node cluster on your laptop you could try Getting Started with Fabric8 and Vagrant which spins up multiple OpenShift V3 boxes configured into a single Kubernetes cluster.
- If you want to try Kuberentes on an operating system that has a JVM with a minimum of fuss then try Getting Started with Fabric8 and Jube which spins up a local Kubernetes cluster using pure Java.
Whichever approach or environment you pick, you’ll be able to reuse the same tools such as the Fabric8 Console and you’ll be able to easily build and run the Quickstarts, view them in the console and connect inside the JVMs of the containers to look at JMX / Camel routes / ActiveMQ queues and so forth using the hawtio console.
We’re still polishing the getting started experience; we hope to keep making it easier and more clear how to get started; along with some videos to show you it in action. (We realise there are a lot of different open source projects involved; so it can appear a little daunting at first!).