Dell Automation Platform: NativeEdge VM fails to obtain DHCP IP from network bridge
Summary: Dell Automation Platform 1.2, Virtual Machines (VMs) fail to obtain an IP address using DHCP when connected to a bridge Virtual Network Segment (VNS) backed by a heterogeneous bond interface (a bond consisting of Ethernet and Wi-Fi). This behavior is the result of a deliberate architectural design choice to ensure broad compatibility with wireless networks. ...
Symptoms
The following observations are seen when encountering the situation:
- A VM is deployed and connected to a bridged Virtual Network Segment using a heterogeneous bond interface (example: Active/Backup bond using
wlp6s0for Wi-Fi andenp3s0for Ethernet) - The VM fails to obtain an IP address using DHCP.
- Inside the VM operating system, the interface link status is UP (
ethtoolshows Link detected: Yes) - NetworkManager logs of the VM (using
journalctl -u NetworkManager -f) reveal DHCPDISCOVERpackets are being sent, but no DHCPOFFERorACKmessages are ever received. - Manually assigning a static IP address to the VM works successfully, allowing external pings and gateway connectivity.
- Using a NAT virtual network segment instead of a Bridge segment successfully assigns an internal/NAT IP address to the VM using DHCP.
Cause
Engineering has confirmed that heterogeneous bond adapters only support the creation of NAT Virtual Network Segments. Bridged Virtual Network Segments over a heterogeneous bond are unsupported.
This is a deliberate design choice made to ensure NativeEdge endpoint successfully and reliably connect to the widest possible variety of third-party Wi-Fi Access Points (APs). The issue stems from the fundamental differences in how standard Ethernet and 802.11 (Wi-Fi) networks handle Layer-2 traffic:
- Wi-Fi Bridging Limitation: The 802.11 standard requires an Access Point to authenticate a single, specific MAC address per client connection. True network bridging allows multiple MAC addresses to pass through a single interface, which Wi-Fi APs typically reject for security reasons.
- Why DHCP Fails: A VM requesting an IP using DHCP does not yet have a routable IP address. It broadcasts a DHCP
DISCOVERpacket at the Data Link Layer (Layer 2) using its own unique virtual MAC address. When the external DHCP server replies with a DHCPOFFER, it targets that virtual MAC. Because that secondary MAC is not authenticated with the Wi-Fi AP, the AP (or the host's forwarding firewall rules) drops the unauthenticated return traffic, breaking the DHCP handshake. - Why a Static IP Works: When assigning a Static IP, the VM bypasses the need for the Layer-2 DHCP broadcast. The hypervisor's networking stack uses Proxy ARP (MAC-NAT) and standard Layer-3 routing. The host intercepts the VM's outbound traffic, routes it on the VM's behalf, masking the traffic behind the host's primary, authenticated MAC address. Because the traffic is properly routed at Layer 3 rather than bridged at Layer 2, network communication flows.
There is no record of this limitation in current documentation, but is added to the Release Notes in the upcoming release.
Resolution
To resolve or work around this issue, use one of the following methods:
- Use a NAT Network Segment: If using a heterogeneous bond adapter (mixed Wi-Fi and Ethernet), configure the virtual network segment to use NAT instead of Bridge.
- Assign a Static IP: If a bridged network connection is strictly required on the heterogeneous bond, assign a static IP address to the VM manually. Static IP configurations bypass the DHCP limitation and allow normal traffic flow.