Stale Transport Nodes in NSX After SDDC Manager Cluster Deletion - VMware Cloud Foundation

Products

• VMware NSX

Issue/Introduction

• When a cluster is deleted from SDDC Manager in a VMware Cloud Foundation (VCF) environment, the following behaviors occur:

  *   The cluster is successfully removed from the SDDC Manager inventory and vCenter Server.
  *   ESXi hosts from the deleted cluster remain visible in the NSX Manager UI.
  *   Orphaned hosts appear under the **Standalone** tab at **System > Fabric > Nodes > Host Transport Nodes**.
  *   Hosts show as **Host Disconnected** or with an **Unknown** status in NSX.

Environment

• VMware NSX-T 3.x
• VMware NSX 4.x
• VMware SDDC Manager 4.4.x
• VMware SDDC Manager 5.x

Cause

The SDDC Manager cluster deletion API uses a cluster-scoped detachment workflow. This workflow removes the Transport Node Profile (TNP) binding and instructs vCenter to destroy the cluster container. Because the cluster entity is removed from vCenter quickly, NSX often loses its parent cluster reference before it can cleanly unprepare and roll back networking on individual hosts. 

This behavior is by design to prevent host network isolation (data path drop/purple screen) during bulk deletion operations. The cleanup process is considered a **best-effort** model; if NSX dependencies are encountered, the decommission proceeds to ensure SDDC Manager inventory consistency. It is a valid behavior. 

Resolution

### Manual Cleanup of Orphaned Hosts

To remove the stale host entries from the NSX fabric:

1. Log in to the NSX Manager UI.
2. Navigate to **System > Fabric > Nodes > Host Transport Nodes**.
3. Change the managed-by dropdown to view standalone hosts or search for the FQDNs of the hosts from the deleted cluster.
4. Select the orphaned hosts.
5. Click **Remove NSX** (Unprepare). 
6. If the hosts no longer exist in vCenter, select the **Force Delete** checkbox.
7. Once the status changes to **Not Prepared**, select the entries again and click **Delete** to purge them from the NSX inventory.

### Scripted Cleanup for Large Volumes

For environments with a large number of stale transport nodes, Broadcom Support provides automation scripts (compatible with NSX 3.2.x and higher). Open a support case and refer to [KB 406768: Script for removing large number of stale transport nodes from NSX]

Additional Information

## Prevention Best Practices

To ensure a clean automated removal in future decommissioning tasks, perform host-scoped cleanup before deleting the cluster container:

1. In SDDC Manager, navigate to the cluster intended for decommissioning.
2. Use the **Remove Host** workflow to remove the ESXi hosts individually or in small batches. This triggers an explicit host-scoped uninstall that NSX can process reliably.
3. Once the cluster is empty, proceed with the **Delete Cluster** workflow in SDDC Manager.

For the latest VCF software and updates, see Download Broadcom products and software for steps to download current releases.