Error "The distributed switch corresponding to this proxy switch is not longer associated with this host" observed while re-adding an ESXi host to cluster

Products

• VMware vCenter Server
• VMware vSphere ESX 7.x
• VMware NSX
• VMware Telco Cloud Platform

Issue/Introduction

Few ESXi hosts were removed from the cluster for hardware maintenance activities and unable to re-add them to the same cluster and observed below error for two VDS configured on the cluster 

Environment

vCenter Server 7.0 Update 3t build 24322018

Vmware ESXi 7.0.3 , 24585291

NSX 3.2.2.0.0.20737185

TCP 2.2

Cause

The ESXi host's local networking configuration has lost synchronization with the vCenter Server database. An orphaned Distributed Virtual Switch (DVS) proxy switch remains active on the host. vCenter Server is prevented from automatically removing this stale proxy switch because the host's management network and physical uplink are still actively bound to it.

Resolution

To resolve this issue, the management network and uplinks must be manually migrated to a standard vSwitch via the ESXi command line. This releases the lock on the stale DVS, allowing it to be removed.

Prerequisites: Ensure out-of-band management access (e.g., ILO, iDRAC, IMM) to the ESXi host console, as network connectivity will drop during these steps.

1. Enable ESXi shell from vcenter.
2. Create a standard switch on the ESXi host from vCenter 
3. Disable coredump network configuration in case enabled (command mentioned below) 
   • esxcli system coredump network get  -- to see if vmk0 is listed as configures network interface and if the service is enabled 
   • esxcli system coredump network set --enable false  --- to disable the service 
4. Access the ESXi host via the Direct Console User Interface (DCUI) ESXi Shell.
5. Identify the stale DVS, active VMkernel ports, and physical uplinks: esxcfg-vswitch -l
6. Remove the existing management VMkernel interface: esxcli network ip interface remove --interface-name=<VMK_ID>
7. Unlink one of the physical uplink (vmnic#) from the stale DVS: esxcfg-vswitch -Q <VMNIC_ID> -V <dvport_ID> <DVS_NAME>
8. Add this physical uplink (vmnic#) to a standard vSwitch: esxcli network vswitch standard uplink add --uplink-name=<VMNIC_ID> --vswitch-name=<STANDARD_VSWITCH_NAME>
9. Recreate the management VMkernel interface and assign it to the standard vSwitch port group: esxcli network ip interface add --interface-name=<VMK_ID> --portgroup-name="<PORTGROUP_NAME>"
10. Reconfigure the IP settings for the management interface: esxcli network ip interface ipv4 set --interface-name=<VMK_ID> --ipv4=<IP_ADDRESS> --netmask=<NETMASK> -g=<GATEWAY_IP> --type=static
11. Verify that the stale DVS has been automatically removed from the host once the dependencies are cleared: esxcfg-vswitch -l
12. From the vSphere Client, migrate the VMkernel adapter and the physical uplink back to the required production Distributed Switch.
13. Migrate the next uplink to the production vDS 
14. Remove the temporary standard switch created from vcenter 

If required enable the coredump network configuration after the ESXi host is successfully added to the cluster 

• • esxcli system coredump network set --enable true -- to enable the service if the target server Ip address and VMkernel interface are configured already.

Additional Information

For any additional information refer KB 431992