Paused behaviour is inconsistent

[enxebre]

What steps did you take and what happened:

In a MachineDeployment:
In the predicates we check cluster.Spec.Paused or the MachineDeployment has the annotation. We ignore MachineDeployment.Spec.Paused https://github.com/kubernetes-sigs/cluster-api/blob/main/internal/controllers/machinedeployment/machinedeployment_controller.go#L75-L97.

Then in the reconciling logic we first check the cluster.Spec.Paused or the MachineDeployment has the annotation. We ignore We ignore MachineDeployment.Spec.Paused https://github.com/kubernetes-sigs/cluster-api/blob/main/internal/controllers/machinedeployment/machinedeployment_controller.go#L126-L130

Then lines below we check only the MachineDeployment.Spec.Paused. We ignore cluster.Spec.Paused or the MachineDeployment has the annotation. https://github.com/kubernetes-sigs/cluster-api/blob/main/internal/controllers/machinedeployment/machinedeployment_controller.go#L225-L227

What did you expect to happen:
Always honour .spec.paused and fallback to the annotation for backward compatibility.
Introduce .spec.paused in MachineSets.
Review all CRDs to make the above consistent.

Anything else you would like to add:
[Miscellaneous information that will assist in solving the issue.]

Environment:

• Cluster-api version:
• minikube/kind version:
• Kubernetes version: (use kubectl version):
• OS (e.g. from /etc/os-release):

/kind bug
[One or more /area label. See https://github.com/kubernetes-sigs/cluster-api/labels?q=area for the list of labels]

[killianmuldoon]

It looks like the current behaviour of paused is different depending on whether you set the spec field or the annotation? With the annotation you get no reconciliation, but with the spec field we end up calling sync() which has even allows some scaling behaviour? 🤔

Is that specific behaviour (sync when paused) used in some of our workflows - maybe a clusterctl move?

[enxebre]

/kind cleanup

[sbueringer]

Q: If we add a paused field to the MachineSet would we then propagate the value of paused from the MD to the MS? If yes, I assume this shouldn't trigger a rollout? (and we have to take care it doesn't specifically)

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.

This bot triages issues and PRs according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue or PR as fresh with /remove-lifecycle stale
• Mark this issue or PR as rotten with /lifecycle rotten
• Close this issue or PR with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[sbueringer]

/remove-lifecycle stale

[fabriziopandini]

/triage accepted