providerID not set automatically by CCM on OCI instances

Hi team,

I'm running a self-hosted Kubernetes cluster on OCI with the latest version of oci-cloud-controller-manager (v1.31.0). I've followed the official instructions, including:

• Setting --cloud-provider=external on all nodes
  

• Setting --provider-id=ocid1.instance... manually on kubelet as a workaround
  

• Using instance principal for authentication (with correct dynamic group policies)
  

• Ensuring instance names match Kubernetes node names
  

• Running the latest CCM version as a DaemonSet
  

However, I noticed that the providerID is not set automatically by the CCM. This leads to the node.cloudprovider.kubernetes.io/uninitialized=true taint not being removed, which prevents scheduling on those nodes until I manually patch them.

In the logs I also see:
error syncing 'node-name': failed to get instance metadata for node: error fetching node by provider ID: compartmentID annotation missing in the node.

This is confusing because I expected the CCM to:

• Automatically detect the node instance ID via IMDS
  

• Automatically set the providerID
  

• Patch compartment-id annotation and remove the uninitialized taint
  

My questions:

• Is CCM supposed to automatically set the providerID field, or is it mandatory to set it manually?
  

• If automatic, what are the required conditions for it to work (e.g. IMDSv1 vs IMDSv2, auth headers, permissions)?
  

• If manual, can this be clarified explicitly in the documentation under the "Preparing Your Cluster" section?
  

Thank you!

Start with the CCM node-sync behavior described in the logs and the official "Preparing Your Cluster" instructions. Determine whether providerID and the compartment-id annotation are expected to be set automatically, then clarify the required setup or manual steps in that documentation; done means the providerID is handled as documented and the uninitialized taint behavior is explained.