vSAN Health Alarm: "MTU check (ping with large packet size)" and Physical NIC Flapping on Broadcom BCM57416

Products

• VMware vSAN

Issue/Introduction

A vSAN cluster reports a "Yellow" health status for the MTU check (ping with large packet size) test. Concurrently, one ESXi host logs indicate intermittent link flapping (Link Down/Up) on physical uplinks, specifically affecting the Broadcom BCM57416 controller.

Similar logs in vobd.log of ESXi host:
[esx.problem.net.dvport.redundancy.lost] Lost uplink redundancy on DVPorts... Physical NIC vmnic# is down.
[vob.net.dvport.uplink.transition.up] Uplink: vmnic# is up.

Environment

VMware vSAN 8.x

Cause

Hardware: Broadcom BCM57416 NetXtreme-E 10GBASE-T RDMA Ethernet Controller
Driver: bnxtnet version 233.0.156.0
Current Firmware is reported as out-of-date
Target Firmware: 23.31.18.10 (Per VMware Compatibility Guide)

The issue is primarily caused by an outdated firmware version on the Broadcom BCM57416 network controllers. This specific firmware level is known to exhibit stability issues that lead to intermittent physical link flapping (vmnic down/up events). When a link flaps or fails to consistently handle large frames during a flapping event, the vSAN MTU health check (which utilizes 9000-byte pings for Jumbo Frame validation) fails, triggering the "Yellow" alarm.

Resolution

To resolve this issue, the firmware for all Broadcom BCM57416 adapters must be updated to the version validated on the VMware Compatibility Guide.

• Confirm the required firmware/driver combination at the Broadcom/VMware Compatibility Guide.
• Schedule a maintenance window and ensure the host is placed into Maintenance Mode with the "Ensure Accessibility" or "Full Data Migration" evacuation policy.
• Apply the latest firmware version to all BCM57416 controllers.
• Perform a cold boot or warm reboot as required by the hardware vendor to initialize the new firmware.