Performance chart for vmnic on the host displays a straight line in VCSA UI

Products

• VMware vSphere ESXi

Issue/Introduction

• Performance chart for NIC on ESXi host might report incorrect values
• Performance chart for vmnic on the host via VCSA UI, the graph might display a straight line even though there is active data transfer
• /var/log/vmkernel.log:

nmlx5_core: vmnic0: nmlx5_en_UplinkWOLStateGet - (nmlx5_core_en_uplink.c:6333) nmlx5_CmdQueryWOLConfig failed: IO was aborted
nmlx5_core: vmnic0: nmlx5_en_UplinkWOLStateGet - (nmlx5_core_en_uplink.c:6347) done, status: Failure
nmlx5_core: vmnic0: nmlx5_en_UplinkWOLStateGet - (nmlx5_core_en_uplink.c:6333) nmlx5_CmdQueryWOLConfig failed: IO was aborted
nmlx5_core: vmnic0: nmlx5_en_UplinkWOLStateGet - (nmlx5_core_en_uplink.c:6347) done, status: Failure
nmlx5_core: vmnic0: nmlx5_en_EcnProtocolQuery - (nmlx5_core_en_ecn.c:247) nmlx5_QueryPortCongStatus failed: 195887328 protocol R_ROCE_RP
nmlx5_core: vmnic0: nmlx5_en_EcnProtocolQuery - (nmlx5_core_en_ecn.c:270) done, status: Failure
nmlx5_core: core: nmlx5_GetECNCap - (nmlx5_core_main.c:281) Fail to query NMLX5_CONG_PROTOCOL_R_ROCE_RP (Failure)

 

Environment

VMware vSphere ESXi 8.0.2

VMware vSphere ESXi 8.0.3

 

Cause

This is known bug in the nmlx5 health mechanism logic where the driver incorrectly detects NIC is in faulty state.

Resolution

This issue is resolved in ESXi 8.0 Update 3e (build number 24674464)

Customer may also update the Mellanox pNIC driver/firmware to a newer version matching the recommendation from Broadcom Compatibility Guide.