The issue is caused by a reconciliation conflict between Tanzu Mission Control Self-Managed and the native vSphere Kubernetes Service (VKS) Add-on Manager. During the upgrade to the 2026.1.21 package repository, both management controllers attempt to update the same PackageInstall custom resource simultaneously, leading to configuration drift and ownership deadlocks.
The cluster-autoscaler must be managed exclusively by Tanzu Mission Control Self-Managed. You must uninstall the existing VKS-managed instance by marking it as unmanaged.
kubectl label cluster addon.addons.kubernetes.vmware.com/cluster-autoscaler=unmanagedPackageInstall status in the guest cluster:kubectl get packageinstall <cluster>-cluster-autoscaler -n vmware-system-tkgThis affects all versions where auto-scaler is supported.