NSX Upgrade Status Resets to NOT_STARTED with Host Version Mismatch Following Upgrade Coordinator Upgrade

Products

• VMware NSX

Issue/Introduction

Following a successful Upgrade Coordinator (UC) upgrade during the VMware NSX upgrade process, the overall NSX upgrade status unexpectedly resets to NOT_STARTED.

As a result:

• Edge Nodes and Management Plane (MP) upgrade states drop to PAUSED.

• Host Nodes status transitions to FAILED with 0% completion.

• A host version mismatch is reported in the UI (e.g., Current Version: 4.1.1.x, Expected Target Version: 4.2.3.x).

• Prior to the Upgrade Coordinator upgrade, the UI confirmed all components had successfully reached 100% SUCCESS.

• Inspecting the Upgrade Coordinator logs (located at /var/log/upgrade-coordinator/upgrade-coordinator.log on the NSX Manager) reveals the following error message:

  No group level upgrade units found... Hence not initializing any progress collectors 

Environment

VMware NSX-T

Cause

This issue occurs because the newly upgraded Upgrade Coordinator fails to map existing upgrade units to their designated inventory groups. Upon restarting into the new version, the UC loses its active tracking groups and the context of previously completed upgrade steps.

Without these tracking groups to reference historical progress, the UC evaluates individual components directly against the new target baseline, triggering false version mismatch failures.

Resolution

1. Restart Manager Services: Log into the NSX Manager CLI and restart the nsx-manager and upgrade-coordinator services to clear the transient in-memory cache and force state rediscovery:

   restart service upgrade-coordinator
   restart service manager
   

2. Verify whether any manual modifications were made to the upgrade plan or node groups during the upgrade process.

3. Resync Upgrade Plan:

   • Refresh the NSX Manager UI Upgrade Plan page.

   • If the status remains stuck after service restarts, re-upload the upgrade bundle metadata to reconcile the state.

Restarting the upgrade-coordinator and manager services clears the corrupted transient in-memory cache created during the UC upgrade. This forces the UC to re-query the NSX database, correctly rediscover existing component states, and align version baselines.

Additional Information

If the issue still persists, please collect the NSX Manager support bundle and  Contact Broadcom support for further assistance.