Steps to Decommission and Remove Tanzu Service Mesh SaaS solution from VKS Clusters
search cancel

Steps to Decommission and Remove Tanzu Service Mesh SaaS solution from VKS Clusters

book

Article ID: 440600

calendar_today

Updated On:

Products

VMware vSphere Kubernetes Service VMware vSphere ESXi

Issue/Introduction

Users need to decommission Tanzu Service Mesh (TSM) components from their Kubernetes clusters. Failure to follow the correct removal sequence results in application pods remaining in "Error" or "Not Ready" states, or a complete inability to schedule new pods cluster-wide.

Environment

  • Tanzu Service Mesh (SaaS)
  • Kubernetes Cluster
  • Tanzu CLI

Cause

Manual deletion of control plane namespaces without prior workload cleanup leaves orphaned istio-proxy sidecars attempting to connect to a defunct SaaS control plane. Furthermore, deleting the control plane without first removing cluster-scoped webhooks will cause the Kubernetes API server to block new pod creations, as it attempts to contact a deleted mutating webhook service.

Resolution

Method A: Formal Cluster Removal (Preferred)

If the SaaS console is accessible, the safest method is to use the official tooling:

  1. Tanzu CLI: Run tanzu service-mesh delete cluster <cluster-name>

  2. TSM UI: Navigate to the Clusters tab, select the target cluster, and click Remove.

Method B: Manual On-Cluster Cleanup

If the console is inaccessible, you must perform a manual cleanup. Contact Broadcom for assistance.

Additional Information

  • Irreversible Action: Operation is irreversible once namespaces and CRDs are deleted. Standard Kubernetes Services must be configured to handle traffic before removal.

  • Feature Loss: Removing TSM terminates mTLS encryption, Global Namespaces (GNS), and mesh-specific traffic splitting rules. 

  • VMware by Broadcom Tanzu Service Mesh Impending End of Life Notice