Kubernetes Cluster on Google Cloud and Expose Couchbase Service

This blog is part of a multi-part blog series that shows how to run your applications on Kubernetes. It will use the Couchbase, an open source NoSQL distributed document database, as the  Docker container.

The first part (Couchbase on Kubernetes) explained how to start the Kubernetes cluster using Vagrant. The second part (Kubernetes on Amazon) explained how run that setup on Amazon Web Services.

This third part will show:

• How to setup and start the Kubernetes cluster on Google Cloud
• Run Docker container in the Kubernetes cluster
• Expose Pod on Kubernetes as Service
• Shutdown the cluster

Here is a quick overview:

Let’s get into details!

Getting Started with Google Compute Engine provide detailed instructions on how to setup Kubernetes on Google Cloud.

Download and Configure Google Cloud SDK

There is a bit of setup required if you’ve never accessed Google Cloud on your machine. This was a bit overwhelming and wish can be simplified.

• Create a billable account on Google Cloud
• Install Google Cloud SDK
• Configure credentials: gcloud auth login
• Create a new Google Cloud project and name it couchbase-on-kubernetes
• Set the project: gcloud config set project couchbase-on-kubernetes
• Set default zone: gcloud config set compute/zone us-central1-a
• Create an instance: gcloud compute instances create example-instance --machine-type n1-standard-1 --image debian-8
• SSH into the instance: gcloud compute ssh example-instance
• Delete the instance: gcloud compute instances delete example-instance

Setup Kubernetes Cluster on Google Cloud

Kubernetes cluster can be created on Google Cloud as:

Make sure KUBERNETES_PROVIDER is either set to gce or not set at all. By default, this provisions a 4 node Kubernetes cluster with one master. This means 5 Virtual Machines are created.

If you downloaded Kubernetes from github.com/kubernetes/kubernetes/releases, then all the values can be changed in cluster/aws/config-default.sh.

Starting Kubernetes on Google Cloud shows the following log. Google Cloud SDK was behaving little weird but taking the defaults seem to work:

There are a couple of unbound variables and a WARNING message, but that didn’t seem to break the script. Google Cloud Console shows:

Five instances are created as shown – one for master node and four for worker nodes.

Run Docker Container in Kubernetes Cluster on Google Cloud

Now that the cluster is up and running, get a list of all the nodes:

It shows four worker nodes. Create a Couchbase pod:

Notice, how the image name can be specified on the CLI. This command creates a Replication Controller with a single pod. The pod uses arungupta/couchbase Docker image that provides a pre-configured Couchbase server. Any Docker image can be specified here.

Get all the RC resources:

This shows the Replication Controller that is created for you.

Get all the Pods:

The output shows the Pod that is created as part of the Replication Controller.

Get more details about the Pod:

Expose Pod on Kubernetes as Service

Now that our pod is running, how do I access the Couchbase server? You need to expose it outside the Kubernetes cluster. The kubectl expose command takes a pod, service or replication controller and expose it as a Kubernetes Service.

Let’s expose the replication controller previously created and expose it:

Get more details about Service:

The Loadbalancer Ingress attribute gives you the IP address of the load balancer that is now publicly accessible. Wait for 3 minutes to let the load balancer settle down. Access it using port 8091 and the login page for Couchbase Web Console shows up:

Enter the credentials as “Administrator” and “password” to see the Web Console:

And so you just accessed your pod outside the Kubernetes cluster.

Shutdown Kubernetes Cluster

Finally, shutdown the cluster using cluster/kube-down.sh script.

Enjoy!