ESXi host experiences CRC errors and link down events

Products

• VMware vSphere ESXi
• VMware Cloud Foundation

Issue/Introduction

• CRC errors reported on specific ESXi host uplinks (e.g., vmnic3).

• vmkernel logs report:

  • Frequent vmnic3: link down notification, vmnic3: link up notification entries

  • link up event received, device running at 0 Mbps so setting queue depth to 0 bytes with expected 0 bytes/us

  • VLANMTUCheck: NMVCRemoveUplink: ####: can't find uplink

• Network traffic is blocked or corrupted on the affected interface.

Environment

ESXi: 8.0U3

VCF: 5.2

Cause

• The issue is caused by a physical or low-level Layer-1 negotiation failure.

• Degraded optical signaling, an unstable DAC twinax connection, or faulty transceivers cause the NIC and switch to fail auto-negotiation.

• This results in the interface temporarily reporting a line speed of 0 Mbps, which forces the ESXi scheduler to set the hardware queue depth to 0, blocking all traffic and causing frame corruption.

Resolution

Address the physical layer components:

• Replace the SFP28/transceiver and the cable connected to the affected uplink (e.g., vmnic3).

• Review the physical switch events to check for Remote Fault, Loss of Signal, or error-disable thresholds triggered by high CRC rates.

• Ensure that Energy Efficient Ethernet (EEE) is explicitly disabled on the switch port, as this can cause intermittent low-voltage synchronization drops.

• Monitor the host logs to ensure the interface stabilizes at its native line speed (e.g., 25 Gbps) and that CRC errors no longer increment.