This project describe how to create a small k8s cluster on azure, and how to deploy a service to internet. In this tutorial, we will use as much as possible docker to not have to install anything else.
Also, this tutorial is planned to have something running in a demo presentation of few hours max. It will not explain every components in details. If you are interested to have a deep understanding, I highly recommend you Kubernetes in action
Finally, this tutorial is not about Docker. Having a base understanding of what are images, containers, and be able to run an image locally is required.
- docker installed locally, using linux container. A base knowledge is required.
- having a azure account with a valid subscription. Take note of the id
- a free account in hub.docker.com
- optionnal : vscode with https://marketplace.visualstudio.com/items?itemName=ms-kubernetes-tools.vscode-kubernetes-tools extension
If you do not wants to install az cli & kubectl cli, you can use this cli to run everything you will need.
From build buildcli folder, build a cli container with docker build -t demo-cli .. From your current terminal, come back to the root starting-k8s directory, then access it with docker run -it -v ${PWD}:/workdir --rm demo-cli. (from powershell, replace ${PWD} by $(PWD)).
Before doing more actions, you need to login to your account, with az login. Follow the instruction by opening url with your browser, and login to your account.
You should be ready to run any azure commands now.
First, you need to create a resource group that will handle your cluster. Simply run az group create --location <location> --name training. e.g.: az group create --location westeurope --name training
The location list can be found by double tapping tab when you already entered az group create --location with white space. If you do not know what to put, you can select westeurope
If you already have a resource group named training, feel free to replace it with another name. We will use training name in all next steps of this tutorial.
Then, you are ready to create your cluster with az aks create -g training -n default-cluster -s Standard_A2_v2 -c 2 --generate-ssh-keys
After a few times, the kubernetes cluster is ready to use ! To configure your kubectl to reach your new cluster, do az aks get-credentials -g training -n default-clusterm
From now, if you want to stop the tutorial to not pay anymore for a small cluster, you can delete it with
az aks delete -g training -n default-cluster. NB: it exists a stop command, but you will still pay a little bit for a service in sleep mode.
I am expecting you to follow the tutorial in the following order.
- ./standalone-service
- ./horizontal-scaling
- ./next if people enjoyed it ..