Removing a NIC from a vDS LAG

Products

• VMware vCenter Server
• VMware vSphere ESXi

Issue/Introduction

This article provides the step by step process for removing a NIC from a vDS LAG to be used as a standalone NIC while both the vCenter and ESXi hosts are still accessible as well as the basic CLI commands to remove the NIC from the vDS.  

LACP (Link Aggregation Control Protocol) Details

• Physical Network
• When the physical network has its NICs configured with LACP enabled these, NICs are essentially bundled together as one. 
• Virtual Network
• The virtual networking will be set up in a similar fashion with the vDS using a LAG (Link Aggregation Group) that contains the NICs that are configured with LACP. 

Note: Any LACP/LAG network troubleshooting will typically require not only support from VMware by Broadcom, but also collaboration with the customer personnel who manages and can access and modify physical switch configurations. Each method of removal listed below will require changes to be made on the physical switch.

           It would be advisable to arrange a maintenance window for the ESXi host where the NIC is set to be removed from the LACP/LAG configuration.

Environment

VMware vCenter server 7.x/8.x

VMware vSphere ESXi 7.x/8.x

Resolution

Removal of a NIC using the vSphere interface:

1. Have the physical network associate access the physical switch configuration and proceed with downing one of the physical NICs that are configured for LACP.
2. Once the physical NIC is down, remove it from the LACP configuration via the physical switch so that it is now a stand alone NIC but do not bring the NIC back up.
3. Confirm which NIC is currently down by either method listed below:
   • Select an ESXi host > Configure > Physical adapters.
     OR
   • Open a putty session to the ESXi host and run the command esxcli network nic list to view which NIC is currently down and verify that the NIC that is down matches the NIC that was downed from the physical network (example output shown below).
      
4. Remove the downed NIC from the LAG. 
   • From the vCenter Network view where the vDS is > Right click the vDS > Select Add and Manage Hosts > Manage host networking > select the ESXi host (or all if this needs to be performed on each ESXi host), by checking the box to the left of its name > Select the NIC that's down and change the Assign uplink to be None > Hit Next until the wizard completes.
5. Verify the NIC is no longer in the LAG by going to the ESXi host > Configure > Virtual Switches > expanding the LAG by clicking the small arrow to the left of the lag.
   • The vmnic# that was removed from the LAG should no longer show. 
6. Have the physical network associate bring the NIC back online
   • NOTE: The NIC will need to remain in a non-LACP configuration on the physical switch while being used as a standalone NIC. Should the NIC be needed back into the LACP configuration, a similar process will need to occur. See Configuring a LAG on a vSphere Distributed Switch Port Group when using LACP.
7. The NIC can now be used as a stand alone NIC with no LACP/LAG configuration.

Removal of a NIC using CLI commands:

1. Have the physical network associate access the physical switch configuration and proceed with downing one of the physical NICs that are configured for LACP.
2. Once the physical NIC is down, remove it from the LACP configuration via the physical switch so that it is now a stand alone NIC but do not bring the NIC back up.
3. Open a putty session to an ESXi host and run the command esxcli network nic list to view which NIC is currently down and verify this output matches the NIC that was downed from the physical network (example output shown above).
4. Locate the DVPort ID of the downed NIC by running the command esxcfg-vswitch -l to list out all virtual switches the ESXi host has access to. Locate the NIC (vmnic) and record the DVPort ID associated with it.
5. Remove the NIC from the vDS by running the command esxcfg-vswitch -Q vmnic# -V <DVPortID> <DVS_Name> (replace DVPortID with the actual ID number and then enter the name of the distributed switch at the end of the command). 
   • Example: The downed vmnic0 needs to be removed from its current vDS, Distributed_Switch, and its currently assigned DVPort ID 513.
     • Command: esxcfg-vswitch -l shows the DVPort ID and the vDS name.
        
       
       
     • Command: esxcfg-vswitch -Q vmnic0 -V 513 Distributed_Switch.
       
   • Have the physical switch associate bring that NIC back up (again, unless the NIC is reconfigured on both the physical switch and virtual network, it will need to remain as a standalone NIC and can only be used as a non-LACP NIC). 
   • The NIC is now off of the vDS and can be used as a standalone NIC.

Additional Information

Below is some documentation that offer possible next steps after removing a NIC from LAG/LACP:

• Recover ESXi host connectivity when management is on DVS
• vCenter network connectivity lost - Recover vCenter network when connected to a Distributed Switch
• Configuring Standard vSwitch (vSS) or virtual Distributed Switch (vDS) from the command line in ESXi
  • ESXCLI Reference
• Preparation for vCenter IP change