Skip to content

Latest commit

 

History

History
56 lines (38 loc) · 1.78 KB

DEVELOPMENT.md

File metadata and controls

56 lines (38 loc) · 1.78 KB
Raw

Build and run kcp

go run ./cmd/kcp start

This will build and run your kcp server, and generate a kubeconfig in .kcp/data/admin.kubeconfig you can use to connect to it:

export KUBECONFIG=.kcp/data/admin.kubeconfig
kubectl api-resources

Build and run Cluster Controller

First, be sure to install the CRD types needed by the controller. These are:

  • Cluster
  • APIResourceImport
  • NegotiatedAPIResource
kubectl apply -f config/

The Cluster Controller requires a --syncer_image to install on new clusters. To build this image and pass it to the Cluster Controller, you can use ko:

go run ./cmd/cluster-controller \
    --syncer_image=$(ko publish ./cmd/syncer) \
    --kubeconfig=.kcp/data/admin.kubeconfig

ko publish requires the KO_DOCKER_REPO env var to be set to the container image registry to push the image to (e.g., KO_DOCKER_REPO=quay.io/my-user).

Test the registration of a Physical Cluster

Registering a physical cluster can be done by simply creating a cluster resource that embeds a kubeconfig file.

For example, in order to register the default cluster of your default kubeconfig file, you can use the following command:

sed -e 's/^/    /' ${HOME}/.kube/config | cat contrib/examples/cluster.yaml - | kubectl apply -f -

Using vscode

Workspace

A configured VSCode workspace is available in contrib/kcp.code-workspace

Debug configuration

You can use the Launch kcp configuration to start the KCP lightweight API Server in debug mode inside VSCode.

If you're using KinD, as the physical cluster you want to register inside KCP, then you can use the Launch cluster controller configuration to debug the cluster controller against a started KCP instance.