DHCP not passing through a SET team

Question

DHCP not passing through a SET team

Chris Higgins 20

i have 2 Dell R640 servers both running Windows Server 2025 and Hyper-V on Broadcom NICs, the first has a VM running DHCP and WDS, if the Hyper-V vswitch is setup using only 1 NIC then VMs on both servers can get DHCP. If i set up the vSwitch with 2 NICs (or 1 NIC but with EmbeddedTeamingEnabled=true) then the VMs on the 2nd server do not get DHCP addresses.

0 comments

Answer accepted by question author

0 additional answers

Your answer

Answer 1

HLBui 13,020 Independent Advisor

Hi Chris Higgins •

When you enable NIC teaming (either by adding two NICs to the vSwitch or setting EmbeddedTeamingEnabled=true), the DHCP traffic sometimes doesn’t get properly passed through because of how the Broadcom drivers handle multicast/broadcast packets. With a single NIC, everything flows fine, but once you team them, the DHCP discover packets from VMs on the second host can get dropped.

The quick fix is usually to check that the Broadcom drivers and firmware are fully up to date older versions are notorious for this. Also, make sure you’re using Hyper‑V’s native switch teaming rather than relying on the vendor’s NIC teaming software. If you want to keep redundancy, configure the vSwitch with SET (Switch Embedded Teaming) but verify that the NICs are in “switch independent” mode and load balancing is set to “dynamic.” That tends to play nicest with DHCP broadcasts.

If you just need stability right now, I’d recommend sticking with a single NIC until you’ve confirmed driver updates and teaming settings. Once that’s sorted, you should be able to run allkeys‑lru (sorry, wrong context meant SET teaming!) without losing DHCP on the second server.

Chris Higgins 20 Reputation points

2026-09-23T10:36:25.5466667+00:00

thanks for your reply, all drivers and firmware are up to date and only using Hyper-V switch teaming.

I really need to have teamed NICs as the 2 servers need to be clustered and resilient. So taking your advice, im going to add in a 3rd box just to host the DHCP VM, without teaming.
HLBui 13,020 Reputation points Independent Advisor

2026-09-23T16:30:55.2433333+00:00

The behavior you’re seeing is tied to how Hyper-V Switch Teaming handles broadcast traffic on Broadcom adapters, and unfortunately DHCP broadcasts are the first casualty. Adding a third box to host the DHCP VM without teaming is a solid workaround, because it takes that critical service out of the teaming path and ensures broadcasts flow properly. Another option you might consider down the line is running DHCP directly on the cluster nodes but binding it to a non‑teamed NIC, so you don’t need extra hardware. Also, keep an eye on Broadcom driver release notes sometimes they quietly fix these broadcast handling quirks in later builds. In the meantime, your plan of isolating DHCP on a standalone host is pragmatic and should stabilize the environment.