Cluster Auto Scaler upgrade fails with error "enforce-node-group-min-size argument should not be passed as an extra argument"

Products

• VMware Tanzu Mission Control - SM
• VMware vSphere Kubernetes Service

Issue/Introduction

After upgrading Tanzu Mission Control - Self Managed from v1.4.2 -> v1.4.4 and Tanzu Standard Package from version 2025.6.18 to 2026.1.21, the Cluster Auto Scaler PackageInstall on guest clusters enters a ReconcileFailed state. While other standard packages reconcile successfully, the Cluster Auto Scaler fails due to a ytt validation error regarding extra arguments.

Symptoms:

• The PackageInstall status shows ReconcileFailed.
• The usefulErrorMessage contains the following snippet: ytt: Error: Evaluating starlark template: - assert.fail: fail: enforce-node-group-min-size argument should not be passed as an extra argument, use enforceNodeGroupMinSize field in arguments
• Attempts to use the suggested enforceNodeGroupMinSize field within extraArguments result in an "unknown flag: - enforceNodeGroupMinSize: true" error.

Environment

• VMware vSphere Kubernetes Service (VMware vSphere Kubernetes Service)
• VMware Tanzu Mission Control - SM (VMware Tanzu Mission Control - SM)

Cause

The underlying cause is a configuration conflict introduced in the upgraded package version. The Cluster Auto Scaler deployment in version 2026.1.21 now includes the --enforce-node-group-min-size flag by default.

If this flag was previously defined in the configuration secret's extraArguments array, the ytt validation triggers a failure to prevent duplicate flags. The error message is misleading because it suggests switching to a camelCase version within the extraArguments array, which is not supported; the flag must be removed from the extraArguments section entirely.

 

Resolution

1. Edit the data-values of autoscaler pkgi secret to ensure the removal of the "enforce-node-group-min-size" line under values.yaml/extraArguments
   • kubectl edit secret -cluster-autoscaler-data-values -n vmware-system-tkg
2. After editing the values, then kick the package to reconcile.
   • kubectl annotate packageinstall -cluster-autoscaler -n vmware-system-tkg packaging.carvel.dev/reconcile-strategy=explicit --overwrite
3. Double check the status
   • kubectl get packageinstall -cluster-autoscaler -n vmware-system-tkg

Note: The functionality provided by this flag is now natively managed by the deployment templates in the 2026.1.21 release. This underlying conflict will be fixed in the next release, eliminating the validation failure.

Additional Information

Subscribe to this knowledge article to get updates on this issue.