Helm is a package manager for Kubernetes that is growing his fame due to semplicity of use, since it basically wraps K8s configurations. In this post I want to give a very short introduction to Helm.
Charts wraps standard K8s Yaml files. Another way of packaging K8s components is operators, which are meant for stateful applications (link 1, link 2).
The CoreOs team shows very well here how an operator works and can be developed.
I will show in a future post how operators can be used on a K8s cluster.
- Install K8s
- Bare Metal On-Premise or Cloud
- Minikube (local)
- Start the K8s cluster
- e.g. minikube start --cpus 8 --memory 8096 --disk-size=20g --vm-driver kvm2
- Install the Helm client - https://docs.helm.sh/using_helm/#installing-helm
- Download version from https://github.com/helm/helm/releases, e.g. "wget https://storage.googleapis.com/kubernetes-helm/helm-v2.12.1-linux-amd64.tar.gz"
- tar -zxvf helm-v2.12.1-linux-amd64.tar.gz
- mv linux-amd64/helm /usr/local/bin/helm
- Install Tiller on the Cluster
- https://docs.helm.sh/using_helm/
- check current context to make sure the correct cluster is referenced (kubectl config current-context)
- initialize helm and install tiller on the cluster
- helm init
- helm init --service-account tiller --kube-context
- Add a Repo and Install Helm Charts
- Please check https://github.com/pilillo/helm-charts for an example
Today I am particularly lazy, so for point 5 I added a full tutorial directly on my Github page.
Charts wraps standard K8s Yaml files. Another way of packaging K8s components is operators, which are meant for stateful applications (link 1, link 2).
The CoreOs team shows very well here how an operator works and can be developed.
I will show in a future post how operators can be used on a K8s cluster.
Enjoy!
Andrea
No comments:
Post a Comment