Removing host from NSX-T transport zone, when the "get logical switch" command provides no details in nsxcli.

Products

• VMware Cloud Foundation

Issue/Introduction

This is to detail how to remove the affected node from the NSX-T manager to allow the decommissioning of the node to complete.

Symptoms:

• Host is part of an NSX-T cluster deployed through SDDC manager.
• In the NSXCLI for this node running the command - get logical switches - returns no data.
• The only method to get the host to resync with NSX-T is to reinstall the host.
• Decommissioning the host from SDDC manager fails at the "Deletion of Transport Node configuration from Host" step.

Resolution

1. On the NSX-T manager get the UUID of the affected node using the following command:

get nodes

2. Find the affected ESXi node from the list.
3. From an SSH session to the SDDC manager run the following command:

curl -k -X DELETE -u 'username:password' 'https://<nsxt-vip-ip>/api/v1/transport-nodes/<transport-node-id> keep_fabric=false&unprepare_host=false&force=true'

 

Where,
'username:password' = is for the root of the NSX Manager.
<nsxt-vip-ip> = Is the virtual IP for the Management Cluster in the Systems > Overview section of the NSX-T client.
<transport-node-id> =  Is the UUID obtained from the get nodes command.

4. Once the command is run, re-try the task from the SDDC Manager UI.