Supervisor TMC-SM Pods Stuck in CrashLoopBackOff Due to Stale LCP Registration

Products

• VMware Tanzu Platform - Kubernetes

Issue/Introduction

A Supervisor cluster exhibits several Tanzu Mission Control (TMC) related pods in a CrashLoopBackOff state.
In the TMC console, the Supervisor status is reported as Unknown, which prevents lifecycle management operations and workload cluster synchronization. Affected pods typically include:

• cluster-health-extension
• extension-updater
• intent-agent
• sync-agent
• tmc-auto-attach
  

All attempts to re-register the Supervisor with Tanzu Mission Control fail 
See: Complete Removal of Tanzu Mission Control Self-Managed (TMC-SM) from vSphere and VCF Environments|
Complete the Registration of a Supervisor

Environment

vSphere with Tanzu 8.x
Tanzu Mission Control 1.4.x

Cause

The Tanzu Mission Control Self-Managed (TMC-SM) database contains a stale Local Control Plane (LCP) registration URL.
This expired or incorrect entry prevents the Supervisor from obtaining a new registration manifest, leading to a mismatch between the Supervisor's local state and the TMC management database.

Resolution

Please open a support case with Broadcom to assist in refreshing the stale registration URL.