The vmconfig-operator and nodeconfig-operator controllers on the management cluster fail to reconcile the NodePolicy for the newly created nodes.
Restarting the vmconfig-operator pod on the management cluster temporarily resolves the issue, allowing the NodePolicy to apply to new nodes.
The NodePolicy status incorrectly displays old nodes that have already been deleted.
TCA 3.4
TCP 5.1
NodePolicy is not supported for use with control plane nodes. Because the functionality is unsupported for this node type, the configuration controllers fail to properly reconcile the policy lifecycle on newly provisioned control-plane nodes.
To achieve the desired configuration goals without using an unsupported NodePolicy, leverage native deployment configurations. An example of some configurations that can be applied without using NodePolicy configurations:
chronyd Configuration: Use the files field to write the chrony.conf content directly to the node. Combine this with postKubeadmCommands to enable and start the chronyd service.
Root CA Trust on Photon OS: Use the files field to place the CA certificate on the node. Add a postKubeadmCommands entry to run the Photon OS certificate trust-refresh step.
SSH Public Keys for CAPV: Utilize the dedicated kubeadmConfigSpec.users field to specify the username and sshAuthorizedKeys directly.