An ESXi host triggers frequent 'Host error' alarms, though the host remains online and operational.
The vmkernel.log file contains repeated alerts regarding the nmlx5_core driver failing to query Wake On LAN (WOL) status.
vmkalert: cpu14:2097549)ALERT: <NMLX_ERR> nmlx5_QueryWol:1057 command failed: IO was aborted
vmkernel: cpu14:2097549)<NMLX_ERR> nmlx5_core: vmnic#: nmlx5_en_UplinkWOLStateGet - (nmlx5_core_en_uplink.c:6350) nmlx5_CmdQueryWOLConfig failed: IO was aborted
Example Alarm Event: Alarm 'Host error' on <hostname> triggered by event 38313955 'Issue detected on <hostname> in <Datacenter>: <NMLX_ERR> nmlx5_QueryWol:1057 command failed: IO was aborted'
VMware vSphere ESXi 7.x
VMware vSphere ESXi 8.x
VMware vSphere ESX 9.x
This issue occurs because the Wake On LAN (WOL) feature is disabled in the server BIOS or at the firmware level for the network adapter. The nmlx5_core driver attempts to query the WOL state (nmlx5_QueryWol), but receives no response or an aborted I/O because the feature is inactive on the hardware.
To resolve the alerts, verify the hardware configuration and enable Wake On LAN in the BIOS.
esxcli network nic get -n vmnic# | grep -i Wakeon
(Replace vmnic# with the interface name reported in the logs, e.g., vmnic4).Supports Wakeon: true
Wakeon: MagicPacket(tm)
This alarm is often benign regarding data traffic stability, as it specifically relates to the driver's inability to read the power management configuration state, there is no performance impact observed due to this.