There are two methods to resolve this inconsistency.
Method 1: UI-Based Cleanup (Recommended)
- Place the impacted ESXi hosts in Maintenance Mode in vSphere.
- Drag the hosts out of the cluster to the Datacenter level. This triggers NSX to un-prepare the hosts.
- In the NSX Manager UI, navigate to System > Fabric > Nodes > Host Transport Nodes.
- Identify any leftover or "phantom" entries for the hosts. Use the Delete or Force Delete option to remove these objects.
- In the NSX UI, verify the Compute Manager connectivity status is "Up".
- Drag the hosts back into the cluster in vSphere.
- Monitor the NSX UI to ensure the hosts successfully transition to a "Success" state.
Method 2: REST API Update If the UI method does not clear the entries, use the NSX API to manually correct the ip_address and display_name fields:
- Retrieve the TN configuration: GET https://<NSX-Manager-IP>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<TN-ID>
- Copy the JSON payload to a text editor and correct the ip_address and display_name.
- Update the TN: PUT https://<NSX-Manager-IP>/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<TN-ID>
Note: For VCF environments, subscribe to this article to receive updates on fix status. Refer to How to subscribe to Knowledge Articles.