Transport Node display name and management IP address appear as duplicates in the NSX Manager UI

Products

• VMware NSX

Issue/Introduction

• When you check in the NSX Manager CLI using the command 'get transport-nodes status' command then the correct non-duplicated names and management IP addresses are returned.
• From the command line on the duplicated hosts using the command 'esxcfg-vmknic -l' you can see that the management vmknic is configured as expected with the correct management IP addresses.
• When you check using an NSX GET API command like the following for both hosts:
  
  GET https://NSX-Manager-IP/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<Transport Node UUID>
  
  then you notice for one of the hosts in the following section of the API output that the "fqdn" field is correct and unique, but the "ip_address" field, which is the management IP for the TN is duplicated:

{
  "node_deployment_info": {
    "fqdn": "<Transport Node Fully Qualified Domain Name>", ----------> Correct and not duplicated
    "ip_addresses": [
      "<Transport Node Management IP>" ----------> Duplicated but is the expected IP for one of the hosts
    ],

and in the following section further down in the output that the "display_name" field is duplicated, and does not match with the "fqdn" field for one of the hosts:

"resource_type": "HostTransportNode",
  "id": "<Transport Node UUID>",
  "display_name": "<Transport Node Display Name>",

• If you check the output from the following Corfu DB browser commands on any of the NSX Managers then you notice the same issues as seen from the API commands:
  
  - /opt/vmware/bin/corfu_tool_runner.py --tool corfu-browser -n nsx -o showTable -t TransportNodeStatus > /tmp/<filename>.txt
- /opt/vmware/bin/corfu_tool_runner.py --tool corfu-browser -n nsx -o showTable -t HostTransportNode > /tmp/<filename>.txt
- /opt/vmware/bin/corfu_tool_runner.py --tool corfu-browser -n nsx -o showTable -t TransportNodeState > /tmp/<filename>.txt

Environment

VMware NSX

Cause

• Using either the REST API or the NSX Manager UI it is possible for someone to update the configuration on a transport node with the same management IP and display name of another existing transport node. 

Resolution

The following workaround assumes TN1 is correct and TN2 was incorrectly updated with a duplicate display name and management IP:

1. Get the Transport Node info for TN2 via the REST API:
   
   GET https://NSX-Manager-IP/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<Transport Node UUID>
   
   
2. Copy the output from step 1 into a text editor, then correct the fields "ip_address" and "display_name" to the expected values for TN2.
   
   
3. Update TN2 using the following API command with the corrected API payload from step 2.
   
   PUT https://NSX-Manager-IP/policy/api/v1/infra/sites/default/enforcement-points/default/host-transport-nodes/<Transport Node UUID>
   
   
4. Verify in the NSX Manager UI and in the REST API using the GET API command from step 1 that the duplication has been removed .

Additional Information

Please refer to the NSX API guide if not familiar with using the REST API:

https://developer.broadcom.com/xapis/nsx-t-data-center-rest-api/latest/api_single_page.html