Link to home
Start Free TrialLog in
Avatar of David Moen
David MoenFlag for Canada

asked on

NIC Teaming Best Practices - Windows Server 2016 Hyper-V

NIC Teaming Best Practices - Windows Server 2016 Hyper-V

I am setting up a new Hyper-V host for a Customer.  It is an HPE ML350 Gen10 with a 4 port network adapter and a dedicated iLO port.  I want to set up a NIC team of course, but I am confused about how to include all the adapters in the team and still be able to give the host OS a fixed IP address.  In the past I have excluded one NIC from the team for the host, but I am reading some blogs that seem to indicate that is a bad idea in terms of redundancy.  On the host I am looking at right now, looking at the network adapters, the only one I can specifically modify the settings of is the one excluded from the team.  Am I missing something?
Avatar of Cliff Galiher
Cliff Galiher
Flag of United States of America image

You can create a team using all NICs.  That'll create a new virtual adapter.

Then in Hyper-v, create an external switch and bind it to that virtual adapter. Check the box to share the adapter with the management OS.

That will create a *new* virtual adapter in the host as well as leave the team virtual adapter in place.

Dont mess with the team virtual adapter. By binding it to the virtual switch hyper-v will manage all that.

Use the new virtual adapter (the second one created when you checked the box on the external virtual switch) to configure your host TCP/IP settings.

That new adapter will go through the same virtue switch as any VMs configured to use it. Which in turn goes through the team VM. Which in turn uses the physical teamed NICs. So you get that redundancy across guests and host.

Otherwise, setup two teams. One team for the host, another for VMs.
Avatar of Philip Elder
Philip Elder
Flag of Canada image

Link to home
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of David Moen


Phil, do you ever sleep?!

Running RemoteApps is probably pretty conservative in terms of bandwidth usage right?  2 x 1GB connections in a team should be tons of pipe for 25-30 RemoteApp users?
Internally, or externally, a 2016+ based RDS single or farm with Win10 on the other end will not consume that much bandwidth at all.

We're running a site plus four external sites with SonicWALL VPN setups across the board to one monolithic app on a Session Host across a 10MbE WiFi based ISP. User count is about 25 -35 across the board. We rarely see bandwidth on the WAN hit anything above 3Mbps.