Supervisor Cluster Unreachable and Host Maintenance Blocked Due to Control Plane Disk Space Full

Products

• VMware vSphere Kubernetes Service

Issue/Introduction

When performing host maintenance or cluster upgrades, the Supervisor cluster may become unhealthy if its control plane nodes run out of disk space. Symptoms include:

• Supervisor control plane nodes showing NotReady
• Control Plane Endpoint (FIP) unreachable
• Cluster operations such as host maintenance or upgrades blocked
• Under Workload Management > Supervisors > Navigate into Cluster > Monitor
  • Configuration Status: Error
  • Kubernetes Status: Error
  • Node Health: Unhealthy

Cause

Supervisor control plane nodes had exhausted their root disk capacity. With no space available, critical services such as kube-apiserver and related components could not function, causing the NSX-T load balancer health checks to fail and the FIP to drop.

Resolution

If the Supervisor cluster is impacted by disk space exhaustion, VMware recommends engaging VMware Support for assistance. Improper manual cleanup of Supervisor node files may cause further disruption to the cluster.

To proceed:

• Collect a vSphere log bundle that includes Supervisor logs
• Open a Support Request and provide the bundle for analysis
• Reference this KB in your case description for quicker triage

Additional Information

• vSphere Supervisor Root Disk Space Full at 100%