openshift / cluster-capi-operator Goto Github PK

the image reference upstreamCAPIOperator needs to be updated. It is starting off as quay.io/asalkeld/openshift-cluster-api-operator-amd64:dev as the operator has not merged upstream yet.

The following branches are being fast-forwarded from the current development branch (main) as placeholders for future releases. No merging is allowed into these release branches until they are unfrozen for production release.

• release-4.16
• release-4.17

For more information, see the branching documentation.

The core cluster controller currently sets conditions on the Cluster object

This goes against Cluster API upstream conventions, given that the condition is owned by the upstream Cluster controller, and Status might go away at any given point in time.

The same behavior can be achieved by using a control plane provider instead that is always initialized.

To ensure that we are not breaking the CI and release payloads, we should have a parallel-e2e, serial-e2e, and a patch-update-e2e, all running as presubmits. This will prevent us from merging code that might disrupt other components of openshift.

Hi all.

I'm trying to integrate Cluster API into my project and I'm a bit confused about this operator compared to the one created by the Kubernetes Cluster API SIG group. That one is based on this proposal and can be found at https://github.com/kubernetes-sigs/cluster-api-operator/. I also noted that https://github.com/openshift/cluster-api-operator is fork of said operator. This is what makes me confused 🤔

The end goal of my project is to let users automatically provision a Kubernetes cluster of their choice on hardware I own (so basically, cluster as a service). I'd like OKD to be one of the choices available. Technically I want to exclusively use Cluster API for the povisioning process, so without Machine API or any other specific tooling. In my search I stumbled upon https://github.com/openshift/enhancements/blob/master/enhancements/machine-api/cluster-api-integration.md, which describes exactly what I need for the OKD part. But that is only the proposal and I'm trying to find if someone already started on the actual implementation.

Any suggestions on repositories I should watch / follow? Or perhaps I'm just looking in totally the wrong place?

For the next version of this tech preview, we should start using ClusterClass, Cluster's managed topologies (which automatically exposes MachineDeployments), and potentially Runtime Hooks.

Upstream docs: https://cluster-api.sigs.k8s.io/tasks/experimental-features/cluster-class/index.html