ESXi host configuration in partial success state with error "VTEP vmk10 failed to be applied, the vnic vmk10 exists"

Products

• VMware NSX

Issue/Introduction

Esxi host stuck in partial success in the NSX configuration with below error on NSX GUI 

Error : Host configuration: VTEP [vmk10,<IP address>] failed to be applied: The vnic vmk10 exists; LogicalSwitch full-sync: LogicalSwitch full-sync realization query skipped.

Environment

NSX : 3.2.0.1.2942942.19682877

ESXi : 7.0.3, 19193900

Cause

Two Distributed virtual switches (DVS) switches were configured on the esxi hosts, The first DVS and its port groups were created for management, vSAN and vmotion vmkernel interfaces and the second DVS was converted as CVDS for NSX segments with LAG interfaces (created from vcenter) for uplinks.

The vmk10 interface got created/configured on the first switch by the system and observed error (as shown in Issue description) while trying to prepare transport node by applying transport node profile on the host.

Resolution

To resolve the issue, perform below actions

1. Set the ESXi host to enter maintenance mode from vcenter.
2. Remove the NSX vibs from the host by taking the host out of the cluster
3. Verify that all NSX vibs are completely uninstalled and reboot the host
4. After the host comes up, the vmkernel interfaces on the host needs to be checked for its presence using command "esxcfg-vmknic -l"
5. If the vmkernel interface (vmk10) is still present on the host, it could be removed using command 

   esxcli network ip interface remove --interface-name=vmk10

6. After removing vmk10 interface, install NSX vibs on the host by bringing it back to the cluster and apply transport node profile on the ESX host from NSX GUI