[Bug] Talos 1.11 to 1.12 upgrade: conflicting hostname configuration (legacy + HostnameConfig) when adding new nodes Provisioning not possible

[marcschaeferger]

After upgrading a Talos cluster vSphere InfraProvider Managed from v1.11.x to v1.12.4, now adding new nodes using a v1.12 machine template results in a configuration validation error:

configuration validation failed:
* stable hostname is already set in v1alpha1 config

The issue appears to be caused by mixed hostname configuration formats (legacy + new HostnameConfig) being present simultaneously.

---

Environment

• Talos Version (cluster): v1.12.4
• Previous Version: v1.11.x
• Platform: VMware / vSphere
• Omni: 1.6.1
• vSphere InfraProvider: v0.1.0-alpha.1-1-g61177a3

---

Problem Description

When adding a new worker node via Omni vSphere InfraProvider (v1.12 template) to an previously upgraded cluster:

• The generated/applied config contains both:

machine:
  features:
    stableHostname: true

and:

apiVersion: v1alpha1
kind: HostnameConfig
...

This leads to a validation error:

stable hostname is already set in v1alpha1 config

---

Observed Behavior

• New node stays in:

  • TYPE: unknown
  • KUBERNETES: n/a
  • KUBELET: n/a

• Stage: Maintenance

• Config is rejected due to validation failure
  See Screenshot

---

Additional Observations

• Existing upgraded nodes (from v1.11):

  • Only contain:

    machine.network.hostname
    machine.features.stableHostname

  • No HostnameConfig present

  • Continue to function correctly

• New nodes (v1.12):

  • Contain both legacy + new hostname config → conflict

---

Logs / Screenshots

---

---

---