An ESXi host utilizing the Intel i40en network driver experiences intermittent link drops and loss of network connectivity. Log analysis reveals repeated transmit queue (TxQ) timeouts and transmit ring disable failures. To recover, the driver initiates a physical function (PF) reset of the network adapter.
Symptoms include:
- The network interface is unable to process transmit requests.
- The driver performs a full hardware reset and rebuilds the internal NIC switch.
- Connectivity typically restores automatically after the reset completes.
Example from /var/run/log/vmkernel.log:
vmkernel: i40en: indrv_UplinkReset:<XXXX_PCI>: <XXXX_VMNIC> : **device reset started** vmkernel: i40en: i40en_UpdateUplinkLinkStatus:<vmkernel: i40en: i40en_PfTxqWait:<XXXX_PCI>: **Tx queue request timeout** vmkwarning: WARNING: i40en: i40en_TxDisableEnableHw:<XXXX_vmkwarning: WARNING: i40en: i40en_SetResetFlags:<XXXX_PCI>vmkernel: i40en: i40en_ResetAndRebuild:<XXXX_vmkernel: i40en: i40en_UpdateUplinkLinkStatus:<
VMware ESXi 8.x
Intel 10Gb Network Adapter (i40en driver)
A broadcast storm or ARP storm in the physical network infrastructure overwhelms the network adapter's processing capacity, leading to driver transmit timeouts.
Coordinate with your network administration team to identify and mitigate the source of broadcast or multicast storms on the physical switch.
Inspect specific VLANs (e.g., VLANs associated with decommissioned or legacy equipment) for excessive traffic.
Verify that Spanning-Tree Protocol (STP) is correctly configured on all physical switch ports connected to ESXi hosts to prevent loops.
Ensure the i40en driver and firmware versions are aligned according to the Broadcom Interoperability Matrix.
For detailed information on troubleshooting spanning tree issues, see Cisco Spanning-Tree Protocol Documentation.
In high-traffic environments, physical network mismanagement is the primary driver for host-level resets. Contact your switch vendor for deep packet analysis if the storm persists.