ClusterctlConfig Resource

The ClusterctlConfig resource allows managing overrides for clusterctl (CAPI Operator) configurations in a declarative way. It is used to configure clusterctl providers and their urls, as well as version restrictions.

ClusterctlConfig follows a GitOps model - the spec fields are declarative user inputs. Turtles does not create or update the resource, it is up to user to specify provider url overrides and maintain its state. It takes precedence over embedded defaults or clusterctl default set of provider definitions.

ADR

Usage

To use the ClusterctlConfig resource:

  1. Create a ClusterctlConfig resource with the clusterctl-config name in the turtles namespace.

  2. The ClusterctlConfig controller will handle updates for the ConfigMap mounted to the cluster-api-operator with the required clusterctl configuration based on the ClusterctlConfig spec.

  3. Manage the ClusterctlConfig object declaratively to apply changes to the generated provider configurations. It may require some time for changes to take effect, as kubelet takes care of updating mounted point based on ConfigMap state.

Here is an example ClusterctlConfig manifest:

apiVersion: turtles-capi.cattle.io/v1alpha1
kind: ClusterctlConfig
metadata:
  name: clusterctl-config
  namespace: rancher-turtles-system
spec:
  providers:
  - name: metal3
    url: https://github.com/metal3-io/cluster-api-provider-metal3/releases/v1.7.1/infrastructure-components.yaml
    type: InfrastructureProvider

This example will generate a clusterctl configuration for the metal3 provider with the specified URL and type.

Specification

The key fields in the ClusterctlConfig spec are:

Deletion

When a ClusterctlConfig resource is deleted, the config map is reverted to its original state, managed by turtles.