Cluster CAPI Operator

The Cluster CAPI Operator manages the installation and lifecycle of the Cluster API Components on Openshift clusters.

Note: This operator only runs on TechPreview clusters.

Managed resources

• CoreProvider - an object that represents core Cluster API and is later reconciled by the upstream operator.
• InfrastructureProvider - an object that represents Cluster API infrastructure provider(AWS, GCP, Azure, etc.) and is later reconciled by the upstream operator.
• Cluster - CAPI Cluster CR that represents current cluster, it is treated as management and workload cluster at the same time.
• InfrastructureCluster - CAPI Infrastructure Cluster CR that represents the infrastructure cluster.
• Worker userdata secret - a secret that contains ignition configuration to be used by the worker nodes.
• Kubeconfig secret - a secret that contains kubeconfig for the cluster.

Controllers

Controllers design can be found here:

• ClusterOperator Controller
• Core cluster Controller
• Infra cluster Controller
• Secret sync Controller
• Kubeconfig Controller

New infrastructure provider onboarding

Steps for infrastructure provider onboarding are documented here.

Running operator locally

Downscale cluster version operator deployment;

kubectl scale deployment cluster-version-operator -nopenshift-cluster-version --replicas=0

Downscale capi-operator deployment:

kubectl scale deployment capi-operator -n openshift-cluster-api-operator --replicas=0

Downscale capi-controllers deployment:

kubectl scale deployment capi-controllers -n openshift-cluster-api --replicas=0

Compile and run operator:

make build && ./bin/capi-operator

Or run the controllers:

make build && ./bin/capi-controllers

Unit tests

make test

Enabling technical preview featureset

kubectl edit featuregate

Set the spec to the following

spec:
  featureSet: TechPreviewNoUpgrade