How to Install ElasticSearch Cluster on K3s/K8s?
This post outlines the process of installing ElasticSearch (ES) on K3s, a lightweight Kubernetes distribution designed
for edge and IoT environments. Before proceeding with the installation, it is crucial to ensure that you have a
prerequisite setup, which includes a working K3s or Kubernetes (K8s) cluster with kubectl connected and configured.
This setup will allow you to effectively manage your cluster's resources and deploy ES using the methodologies described
here. The following steps will guide you through the installation of the ElasticSearch cluster, leveraging the
Kubernetes operator and Helm charts for a simplified and efficient deployment.
Note
Kubernetes uses namespaces to separate different projects. We will use separate namespaces for each Tracardi dependency.
The best way to install ElasticSearch (ES) on Kubernetes is to use the ES operator. A Kubernetes operator is a method of packaging, deploying, and managing a Kubernetes application. To install it, type:
kubectl apply -f https://download.elastic.co/downloads/eck/2.10.0/crds.yaml
kubectl apply -f https://download.elastic.co/downloads/eck/2.10.0/operator.yaml
It will automatically install the ElasticSearch Kubernetes Operator, which configures and manages ES clusters in your K8s environment.
Helm Chart Download
Then we will use a helm chart prepared by Tracardi to simplify the whole installation. First, download the helm chart file:
This command saves the helm chart as a file named elastic-1.0.0.tgz.
Configuration
Next, you need to prepare a simple file called values.yaml that defines how many nodes your Elastic cluster should
have:
version: 8.11.3 # Version of ElasticSearch
service:
type: LoadBalancer # LoadBalancer for ES
port: 9200
master:
replicas: 1 # Number of master nodes
storage:
size: 5Gi # Storage size for the master node
data:
replicas: 3 # Number of worker nodes
storage:
size: 10Gi # Storage size for each worker node
This is a very simplified installation but will install a production-ready ElasticSearch cluster. It will create storage volumes and the defined number of nodes.
Installation
Let's install ElasticSearch. We will use the previously defined values.yaml file and the downloaded helm chart:
This command installs the ElasticSearch cluster using the helm chart elastic-1.0.0.tgz within the namespace elastic,
applying your configuration specified in values.yaml. This automates the setup of your ES cluster with specified
versions, node configurations, and storage settings.
If later you would like to upgrade your installation after editing values.yaml, please use the following command. Note
that you will not be able to decrease the storage; you can only increase it.
Wrap Up
After a minute or two, you should be able to see the pods in the elastic namespace. A pod in Kubernetes is the
smallest deployable unit that can be created, scheduled, and managed. It's essentially a group of one or more docker
containers, with shared storage/network, and a specification for how to run the containers.
To check if everything is OK, run this command: