Kubernetes API Access¶

The kubernetes api can be accessed by a utility called kubectl and our clusters are integrated with the UCAR user authentication. These utilities are well documented online so we will point to external documentation for most of this and provide additional instructions as necessary.

Install kubectl¶

See https://kubernetes.io/docs/tasks/tools/#kubectl

Install Azure kubelogin¶

See https://azure.github.io/kubelogin/install.html

Set up your kubernetes config file¶

kubectl uses either ~/.kube/config by default or the KUBECONFIG environment variable to know which file to load its configuration from. We recommend using the default file location.

The CIRRUS team will provide you with the necessary configuration file.

Login¶

Once the utilities are installed correctly when you use kubectl it will start the login process by asking you to go to a microsoft/azure url and what account to log in with. It also uses two factor authentication.

Extra Tips¶

The kubectl quick reference is very useful https://kubernetes.io/docs/reference/kubectl/quick-reference/

Enable shell completion for the kubectl command. This will allow tab completing resources. https://kubernetes.io/docs/reference/kubectl/quick-reference/#kubectl-autocomplete

You will be typing kubectl a lot so alias it to k. Make sure to enable bash completion for the alias. This should be put in ~/.bashrc https://kubernetes.io/docs/reference/kubectl/quick-reference/#bash

Switching clusters and namespaces may also happen frequently. To make this easier install the kubectx and kubens utilities. https://github.com/ahmetb/kubectx

Because you are switching clusters and namespaces it is easy to lose track of where you are. To help with this install kube-ps1 to modify your shell with the context and namespace you are currently working in. https://github.com/jonmosco/kube-ps1