GitHub - GeertJohan/kube-multi-config

kube-multi-config

A simple, secure and effective method to manage kube config files. Without installing anything.

Some backstory:

My .kube/config file contained hundreds of lines of yaml. It became a pain to edit the file. I wanted to simplify this without installing any tools to manage the file. It should be possible to add/remove kubernetes contexts manually.

This walkthrough helps you setup a simple structure to manage large number of kubernetes contexts.

How it works

The KUBECONFIG environment variable allows kubectl users to list individual kube config files, separated by a semicolon. This is the feature that you'll use to separate the single .kube/config file into multiple config files. Each config file can still contain more than one cluster/context/user, it's up to you how to organize these.

A script will automatically update the KUBECONFIG environment variable so that you don't have to synchronize the list of configs and the KUBECONFIG variable pointing to those configs.

Setup

Lets start by creating some directories

mkdir ~/.kube-multi-config
# The config.d folder will contain the kubeconfig files.
mkdir ~/.kube-multi-config/config.d

Next, let's create a bash/zsh compatible script to compose the KUBECONFIG environment variable.

nano ~/.kube-multi-config/export-kubeconfig

(Or use your favorite editor to create this file.)

Add this code:

# List of files in config.d
kubeConfigFileList=$(find ${HOME}/.kube-multi-config/config.d -type f)

# Always include the original kubeconfig file for backwards compatibility. Also,
# some tools (for example `gcloud`) add kube contexts directly to
# `.kube/config`.
KUBECONFIG="${HOME}/.kube/config"

# Combine all file paths into the single `KUBECONFIG` variable.
while IFS= read -r kubeConfigFile; do
  KUBECONFIG+=":${kubeConfigFile}"
done <<< ${kubeConfigFileList}

# Export the `KUBECONFIG` variable so that it can be 'seen' by commands
# following this script.
export KUBECONFIG

Note that the file doesn't have to be executable.

To run this script every time you open a terminal/shell, add it to ~/.bashrc or ~/.zshrc (depending on which shell you use) using the source builtin.

source ${HOME}/.kube-multi-config/export-kubeconfig

Now add a config file to the config.d folder.

cd ~/.kube-multi-config/config.d
touch my-project
chmod 600 my-project
nano my-project

Add a config file, it could look something like this:

apiVersion: v1
kind: Config

clusters:
- name: my-project_cluster
  cluster:
    certificate-authority-data: <ca blob>
    server: https://<server address>

users:
- name: my-project_admin
  user:
    token: <secret token>

contexts:
- name: my-project
  context:
    cluster: my-project_cluster
    user: my-project_admin

Feel free to copy groups of cluster/context/user enries from ~/.kube/config files into separate files in ~/.kube-multi-config/config.d.

Note that the ~/.kube/config needs to remain. It's the only file that contains the current-cluster: line.

Restart your shell and echo $KUBECONFIG to see the new config files. Run kubectl config get-contexts to list all your contexts. You can still use kubectl config use-context <your context name> to switch to any context in any config file.

License: This walkthrough and the scripts are license under a BSD license