LACP Negotiation Failure (AGG,DEF) on ESXi 8.0

Products

VMware vSphere ESXi

Issue/Introduction

ESXi hosts utilizing any network adapters may experience intermittent or partial Link Aggregation Control Protocol (LACP) negotiation failures on specific physical uplinks. While the physical link state appears as Up, the port fails to bundle correctly within a Link Aggregation Group (LAG).

Symptoms include:

Uplink port state stuck in AGG,DEF (Defaulted) instead of the expected AGG,SYN,COL,DIST (Bundled).
Significant asymmetry in LACPDU counters where the host transmits successfully (Tx) but fails to receive responses (Rx) reliably.
Intermittent loss of Cisco Discovery Protocol (CDP) or Link Layer Discovery Protocol (LLDP) information on the affected port.
The issue persists even when the NIC is moved from a Distributed Virtual Switch (DVS) to a standard vSwitch.

Environment

Primary Product: VMware vSphere ESXi 8.0 Update 3 (and later)
Configuration: vSphere Distributed Switch (VDS) with LACP/LAG configured.

Cause

The root cause is typically a hardware-level receiver fault isolated to a specific port on the network adapter, or a failure of the physical SFP/transceiver and associated fiber cabling.

While symptoms like LACPDU asymmetry can sometimes be associated with firmware errata, the fault is isolated to hardware when the issue persists after updating to known-stable driver/firmware baselines and after normalizing all software configurations.

Resolution

Follow these validated steps to isolate and resolve the physical layer failure:

Verify LACP Statistics: Execute the following command in the ESXi Shell to confirm LACPDU asymmetry:
```
localcli network vswitch dvs vmware lacp stats get
```
Identify if Rx LACPDUs remain static or significantly lower than Tx LACPDUs.
Rule out Software/Firmware Variables: Ensure the affected host is running the latest recommended driver and firmware for the vmnic using the Broadcom Compatibility Guide: https://compatibilityguide.broadcom.com/
Physical Layer Isolation: If the issue is isolated to a single port while other ports on the same card function correctly:
- Reseat the SFP/Transceiver and the fiber cable on both the host and switch side.
- Swap the SFP and fiber cable with a known-working port to see if the fault follows the hardware.
Hardware Replacement: If symptoms persist after swapping cables/SFPs, replace the physical SFP and fiber cable associated with the faulty vmnic. In many cases, a hardware-level investigation by the vendor (e.g., HPE) may be required to authorize an adapter replacement.

Additional Information

To speak with a customer representative or a Support Engineer see Contact Support.