Nodes do not join the cluster

[clickersmudge]

Hi,
I have a KubeOne cluster v1.7.3

kubectl get machinedeployment -A
NAMESPACE     NAME              REPLICAS   AVAILABLE-REPLICAS   PROVIDER   OS       KUBELET   AGE
kube-system   prod-pool1        0                               hetzner    ubuntu   1.27.11   12d
kube-system   prod-worker-arm   1          1                    hetzner    ubuntu   1.27.11   12d

When it adds a new machinedeploymnet from a file that worked early on.
The machine is created at Hetzner.

A new machinedeployment is created

but a new node never appears

kubectl  describe machinedeployment -n kube-system <NAME>

Shows warning.
Warning ReconcileError 2m42s (x3 over 2m42s) machinedeployment-controller Operation cannot be fulfilled on machinedeployments.cluster.k8s.io "prod-worker-x86": the object has been modified; please apply your changes to the latest version and try again

I checked further

kubectl get machines --namespace kube-system
kubectl get machineset --namespace kube-system

I created a new MachineDeployment with a different name

The machine has also created.

After about 10-15 minutes, the old machine is deleted, a new one is created, but none joins the cluster.

 kubectl logs -n kube-system                     machine-controller-7b47d7b45b-bhf5h

[kron4eg]

Hello!

Yeah, we are aware of the situation (see #3165) and already merged the fix, the releases is pending. For now you can fix this by deploying updating version of the OSM (see #3163 for 1.7 release).