NSX configuration failure on ESXi transport node after motherboard replacement.

Products

• VMware NSX

Issue/Introduction

After replacing a motherboard on an ESXi host that is part of an NSX-T/NSX 4.x environment, the host may show a status of "Install Failed," "Disconnected," or "Unreachable" in the NSX Manager UI. Attempts to "Configure NSX" or "Resolve" the state fail with errors indicating the host cannot be reached or is in Maintenance Mode. Additionally, the standard "Remove NSX" option may be greyed out or fail to execute.

Environment

VMware NSX 4.x
VMware ESXi 7.x and 8.x
VMware Cloud Foundation (VCF)

 

Cause

NSX identifies Transport Nodes using a unique management plane identity often tied to the host's BIOS UUID. A motherboard replacement changes this UUID, causing a mismatch between the host's actual identity and the record stored in the NSX Manager database. Because the Manager is trying to communicate with the "old" identity, it cannot successfully push configuration updates or perform a standard uninstall.

Resolution

To resolve this, the orphaned record must be forcefully removed from the NSX Manager and the host must be locally cleaned before re-preparation.

1. Perform Force Delete in NSX Manager:

1. Log in to the NSX Manager UI.
2. Navigate to System > Fabric > Nodes > Host Transport Nodes.
3. Locate the affected host (if not in its cluster, check the Standalone or Other Nodes tab).
4. Select the host and click Delete or Remove NSX.
5. In the confirmation dialog, you must select the Force Delete checkbox and click Delete. This clears the entry from the NSX database without requiring communication with the host.

2. Manual Local Cleanup:

1. Ensure the ESXi host is in Maintenance Mode.
2. Access the host CLI via SSH as root.
3. Run the following command to remove local NSX components: 

   nsxcli -C del nsx

4. Verify that NSX VIBs are removed by running: 

   esxcli software vib list | grep nsx

3. Re-prepare the Host:

1. Reboot the ESXi host to ensure all stale management processes are cleared.
2. In vCenter, move the host back into the NSX-prepared cluster.
3. The NSX Manager will detect the host as a "new" node and automatically begin the installation process.

Additional Information

If the host remains in a "Disconnected" state after these steps, verify that the management plane connectivity (ports 1234, 1235) is not being blocked by a firewall and that the Compute Manager (vCenter) status is "Up" in the NSX Manager.

Uninstalling NSX from a Host Transport Node