Chekkan's Blog

Setting up Elasticsearch cluster on Kubernetes - Part 1 - Single Node Cluster

January 11, 2018 - 4 min read

This is the begining of a multiple part series of blog posts going through setting up Elasticsearch cluster in kubernetes.

The main aim of this series of blog posts will be make notes for myself as I try to learn kubernetes and for anyone in the same position.

In this blog post, I will just concentrate on useful Kubernetes getting started resources, commands, and also with an aim of creating a single node Elasticsearch cluster.

Getting Started

I found the most helpful resource for me was the Kubernetes official website for starting to learn kubernetes. Head over to the Interactive Tutorials section of the website and spend time going through all 6 modules to cover the basics. It should only take you 1 - 2 hours. You won’t have to install anything to try it out.

The next step I took was followed the Hello Minikube tutorial. This helped me to get minikube and kubectl commands installed. (Minikube is the local development Kubernetes environment and kubectl is the command line interface used to interact with Kubernetest cluster).

Shaving the Yak!

One or two commands that I use in this post will be mac specific. Reference this guide to get more up to date and OS specific commands.

Once you’ve got the tools all installed, you can now follow along these steps to create a single node Elasticsearch cluster.

If you are using Minikube, make sure that its started properly by running this command (for mac):

minikube start --vm-driver=hyperkit

Now set the Minikube context. The context is what determines which cluster kubectl is interacting with.

kubectl config use-context minikube

Verify that kubectl is configured to communicate with your cluster:

kubectl cluster-info

To view the nodes in the cluster, run:

kubectl get nodes

Kubernetes Dashboard

Minikube includes the kubernetes dashboard as an addon which you can enable.

minikube addons list

returns

- default-storageclass: enabled
- coredns: disabled
- kube-dns: enabled
- ingress: disabled
- registry: disabled
- registry-creds: disabled
- addon-manager: enabled
- dashboard: enabled
- storage-provisioner: enabled
- heapster: disabled
- efk: disabled

You can enable an addon using:

minikube addons enable dashboard

You can then open the dashboard with command

minikube addons open dashboard

Screen-Shot-2018-01-24-at-15.49.17

Single Node Elasticsearch Cluster

Let’s start off by creating a single node elasticsearch cluster. According to the elasticsearch documentation, the current version at the time of this writing is 6.1.1. And there are three flavours of docker images. We will just use the basic image which has xpack and free license.

Run the following command to deploy elasticsearch container into our kubernetes environment exposing just the port 9200. There is no way to expose multiple ports using the kubectl command line currently. We will probably revisit this in a later post.

kubectl run elasticsearch --image=docker.elastic.co/elasticsearch/elasticsearch:6.2.1 --env="discovery.type=single-node" --port=9200

To list your deployments use:

kubectl get deployments

To list all pods and watch the container getting created:

kubectl get pods --watch

Save the pod name as a variable for use in later commands.

export POD_NAME=$(kubectl get pods -o go-template --template '{{range .items}}{{.metadata.name}}{{"\n"}}{{end}}')
echo Name of the Pod: $POD_NAME

Note: The above command will not work if you have multiple pods.

Now, you can take a look at the logs using the command:

kubectl logs $POD_NAME

Expose the cluster

We can verify that the cluster is running by looking at the logs. But, let’s check if elasticsearch api is responding first.

In a seperate shell window, excute the following to start a proxy into Kubernetest cluster.

kubectl proxy

Outputs:

Starting to serve on 127.0.0.1:8001

Now, back in the other window, lets execute a curl command to get the response from the pod via the proxy.

curl http://localhost:8001/api/v1/proxy/namespaces/default/pods/$POD_NAME/

Update - 18 Feb 2019: In later version of kubernetes, the proxy url is at http://localhost:8080/api/v1/namespaces/default/pods/$POD_NAME/proxy/

Outputs:

{
  "name": "xNAF2Lj",
  "cluster_name": "docker-cluster",
  "cluster_uuid": "sD73f3CkQJOX3t4sxraSug",
  "version": {
    "number": "6.2.1",
    "build_hash": "7299dc3",
    "build_date": "2018-02-07T19:34:26.990113Z",
    "build_snapshot": false,
    "lucene_version": "7.2.1",
    "minimum_wire_compatibility_version": "5.6.0",
    "minimum_index_compatibility_version": "5.0.0"
  },
  "tagline": "You Know, for Search"
}

Great, everything is working.

Now, lets expose this deployment to outside of Kubernetes network:

kubectl expose deployment elasticsearch --type=LoadBalancer

Pro tip Use MiniKube to open the service in your default browser.

minikube service elasticsearch

In my case, the port that was assigned to this pod was 31389. But, we have elasticsearch cluster now running in Kubernetes!


Harish H. Babu

Written by Harish H. Babu who lives and works in Cardiff engineering software things. You should follow him on Twitter