An administrator attempts to upgrade an existing Tanzu Kubernetes Cluster (TKC).

Log in to the Supervisor as an administrator.

Run kubectl to look up the updatejob_name.

Run kubectl proxy so that curl can be used to issue requests.

Use curl to issue a patch command to clear the .status.conditions on the Job controller.

In VMware Cloud Foundation (VCF) 9.0, managing Tanzu Kubernetes Clusters (TKC) through the vSphere Kubernetes Service (VKS) involves a declarative lifecycle managed by the vSphere Supervisor. When a TKC upgrade enters an upgradefailed state, it is typically because the underlying cluster reconciliation logic has hit a terminal error or a timeout within the Cluster API (CAPI) controllers. To resolve this, an administrator must interact directly with the Supervisor’s API server to reset the state of the upgrade job.

The troubleshooting workflow begins with authenticating to the Supervisor cluster using the kubectl vsphere login command (Step B). Once authenticated and the context is set to the correct namespace, the administrator must identify the specific job responsible for the upgrade failure by querying the namespace for job resources (Step D). Because the Supervisor API requires specific authentication tokens that are complex to manage via standard CLI tools for manual JSON patching, running kubectl proxy (Step A) is the standard procedure to create a local gateway. This allows the administrator to use a simple curl command (Step C) to send a JSON Merge Patch to the API server. The critical step is clearing the .status.conditions field. By removing the " Failed " condition, the Supervisor ' s controller-manager is triggered to re-evaluate the desired state and attempt the upgrade process again. This sequence ensures the declarative engine can resume operations without requiring a full cluster redeployment.