Deploying applications on Kubernetes can often be a complex process. Helm charts simplify this by creating a package of pre-configured Kubernetes resources. In this post, we're going to create a Helm chart to deploy a production-ready etcd cluster.
What is a Helm Chart?
A Helm chart is a collection of files that describe a related set of Kubernetes resources. A single chart might be used to deploy something simple, like a memcached pod, or something complex, like a full web application stack with HTTP servers, databases, caches, and so on.
Prerequisites
Installation
First, download Helm's installation script:
$ curl https://raw.githubusercontent.com/helm/helm/master/scripts/get-helm-3 > get_helm.sh
Next, give the script executable permissions:
$ chmod 700 get_helm.sh
Finally, run the installation script:
$ ./get_helm.sh
Setting Up the Helm Chart
Creating a Helm chart involves creating a specific directory structure with several different files. Here's a simple outline of our etcd Helm chart:
etcd/
├── Chart.yaml
├── templates/
│ ├── _helpers.tpl
│ ├── service.yaml
│ ├── statefulset.yaml
│ └── pvc.yaml
└── values.yaml
Understanding the Chart Files
Chart.yaml: This is where we define the chart's metadata.
Here's an example:
apiVersion: v3
name: etcd
version: 0.1.0
description: A Helm chart for deploying an etcd cluster in Kubernetes
templates/: This directory contains the template files that generate Kubernetes manifest files when the chart is installed.
_helpers.tpl: This is a place to put template helpers that you can reuse throughout your chart.
{{/* Generate the name */}}
{{- define "etcd.fullname" -}}
{{- .Chart.Name }}-{{ .Release.Name }}
{{- end -}}
service.yaml: This file defines a Service, which can be used to access your etcd cluster from other applications in your Kubernetes cluster.
apiVersion: v1
kind: Service
metadata:
name: {{ include "etcd.fullname" . }}
spec:
clusterIP: None
ports:
- port: 2379
name: client
- port: 2380
name: peer
selector:
app: etcd
statefulset.yaml: This file defines a StatefulSet, which is used to manage the etcd Pods.
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: {{ include "etcd.fullname" . }}
spec:
serviceName: "{{ include "etcd.fullname" . }}"
replicas: {{ .Values.replicaCount }}
selector:
matchLabels:
app: etcd
template:
metadata:
labels:
app: etcd
spec:
containers:
- name: etcd
image: "{{ .Values.image.repository }}:{{ .Values.image.tag }}"
ports:
- containerPort: 2379
name: client
- containerPort: 2380
name: peer
volumeMounts:
- name: data
mountPath: /var/run/etcd
volumeClaimTemplates:
- metadata:
name: data
spec:
accessModes: [ "ReadWriteOnce" ]
storageClassName: {{ .Values.storageClass }}
resources:
requests:
storage: {{ .Values.storageSize }}
pvc.yaml: Persistent Volume Claim (PVC) for storage of etcd data.
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: data
spec:
accessModes:
- ReadWriteOnce
storageClassName: {{ .Values.storageClass }}
resources:
requests:
storage: {{ .Values.storageSize }}
values.yaml: This is where you define the default values for your chart. Here's a simple example:
replicaCount: 3
image:
repository: quay.io/coreos/etcd
tag: v3.5.0
storageClass: standard
storageSize: 1Gi
Run your etcd Helm chart
Make sure you are in the proper kubectl context and create a namespace named etcd.
$ kubectl create namespace etcd
Navigate to the directory containing the Chart files
$ helm install etcd-cluster . --namespace etcd
Output:
NAME: etcd-cluster
LAST DEPLOYED: Sat Aug 5 23:15:25 2023
NAMESPACE: etcd
STATUS: deployed
REVISION: 1
TEST SUITE: None
Get the resources
$ kubectl get all -n etcd
Output:
NAME READY STATUS RESTARTS AGE
pod/etcd-etcd-cluster-0 1/1 Running 0 104s
pod/etcd-etcd-cluster-1 1/1 Running 0 96s
pod/etcd-etcd-cluster-2 1/1 Running 0 91s
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/etcd-etcd-cluster ClusterIP None <none> 2379/TCP,2380/TCP 104s
NAME READY AGE
statefulset.apps/etcd-etcd-cluster 3/3 104s
Conclusion
This Helm chart sets up an etcd cluster on a Kubernetes environment, using best practices such as StatefulSets for pod identity and PVCs for data persistence.
This setup, however, is quite basic and might need to be adjusted based on your specific use case or production environment. Always remember to test your Helm charts thoroughly in a staging environment before deploying them in production.
Top comments (0)