Quick Start with Docker provider generating 9 (nine) nodes cluster instead of 6 (six)

[dmvolod]

What steps did you take and what happened?

Going through Quick Start with Docker and got following results

kubectl get machines
NAME                                     CLUSTER           NODENAME                                 PROVIDERID                                          PHASE     AGE   VERSION
capi-quickstart-7bvjq-7ddxm              capi-quickstart   capi-quickstart-7bvjq-7ddxm              docker:////capi-quickstart-7bvjq-7ddxm              Running   18m   v1.32.0
capi-quickstart-7bvjq-czg5m              capi-quickstart   capi-quickstart-7bvjq-czg5m              docker:////capi-quickstart-7bvjq-czg5m              Running   17m   v1.32.0
capi-quickstart-7bvjq-rjpmm              capi-quickstart   capi-quickstart-7bvjq-rjpmm              docker:////capi-quickstart-7bvjq-rjpmm              Running   11m   v1.32.0
capi-quickstart-md-0-bvhfn-kr69p-2j4t8   capi-quickstart   capi-quickstart-md-0-bvhfn-kr69p-2j4t8   docker:////capi-quickstart-md-0-bvhfn-kr69p-2j4t8   Running   18m   v1.32.0
capi-quickstart-md-0-bvhfn-kr69p-k8zb9   capi-quickstart   capi-quickstart-md-0-bvhfn-kr69p-k8zb9   docker:////capi-quickstart-md-0-bvhfn-kr69p-k8zb9   Running   18m   v1.32.0
capi-quickstart-md-0-bvhfn-kr69p-q27mz   capi-quickstart   capi-quickstart-md-0-bvhfn-kr69p-q27mz   docker:////capi-quickstart-md-0-bvhfn-kr69p-q27mz   Running   18m   v1.32.0
worker-3dkl6x                            capi-quickstart   capi-quickstart-worker-3dkl6x            docker:////capi-quickstart-worker-3dkl6x            Running   17m   v1.32.0
worker-3fq5r4                            capi-quickstart   capi-quickstart-worker-3fq5r4            docker:////capi-quickstart-worker-3fq5r4            Running   17m   v1.32.0
worker-uw8gfz                            capi-quickstart   capi-quickstart-worker-uw8gfz            docker:////capi-quickstart-worker-uw8gfz            Running   17m   v1.32.0

kubectl --kubeconfig capi-quickstart.kubeconfig get nodes
NAME                                     STATUS   ROLES           AGE     VERSION
capi-quickstart-7bvjq-7ddxm              Ready    control-plane   10m     v1.32.0
capi-quickstart-7bvjq-czg5m              Ready    control-plane   9m45s   v1.32.0
capi-quickstart-7bvjq-rjpmm              Ready    control-plane   3m49s   v1.32.0
capi-quickstart-md-0-bvhfn-kr69p-2j4t8   Ready    <none>          10m     v1.32.0
capi-quickstart-md-0-bvhfn-kr69p-k8zb9   Ready    <none>          9m53s   v1.32.0
capi-quickstart-md-0-bvhfn-kr69p-q27mz   Ready    <none>          10m     v1.32.0
capi-quickstart-worker-3dkl6x            Ready    <none>          9m28s   v1.32.0
capi-quickstart-worker-3fq5r4            Ready    <none>          9m28s   v1.32.0
capi-quickstart-worker-uw8gfz            Ready    <none>          9m28s   v1.32.0

It seems to clusterctl generate logic has a bug for Docker provider.

What did you expect to happen?

Expected to generate 6 nodes cluster as described in the guide.

Cluster API version

clusterctl version
clusterctl version: &version.Info{Major:"1", Minor:"9", GitVersion:"v1.9.5", GitCommit:"068c0f3c8ed3a7e382952f8ae68b648aee240a60", GitTreeState:"clean", BuildDate:"2025-02-18T19:07:25Z", GoVersion:"go1.22.12", Compiler:"gc", Platform:"linux/amd64"}

Kubernetes version

kubectl version
Client Version: v1.29.14
Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
Server Version: v1.32.2

Anything else you would like to add?

No response

Label(s) to be applied

/kind bug
/area provider/infrastructure-docker

[chrischdi]

The issue on this is currently the flag for the number of workers in clusterctl and its value gets used in the template to set:

• number of replicas for MachineDeployment
• number of replicas for MachinePools

We could split up the template to have separate ones which don't contain the one or the other.

/priority backlog
/triage accepted

/assign @sbueringer

[chrischdi]

And thanks for reporting this @dmvolod 🎉