VM lost network connectivity due to I350 Nic Rx Hang

Products

• VMware vSphere ESXi

Issue/Introduction

• VMs mapping on one Nic card (vmnic0 for example) lost network connection.
• VM could not resolve gateway ARP due to failing to receive gateway ARP reply.
• Use pktcap-uw command to do packet capture on ESXi, vmnic0 does not receive ARP reply from gateway.
• Nic hardware is Intel I350 Gigabit Network card, with igbn driver 1.4.11.7. 
• vmkernel.log has repeated messages "igbn: igbn_CheckRxHang: vmnic#: false hang detected on RX queue #"
• Down and up vmnic0 finally solved the problem.

Environment

VMware vSphere ESXi

Cause

The log entries "igbn: igbn_CheckRxHang: vmnic#: false hang detected on RX queue #" indicate that the network driver has identified an internal processing stall within the hardware's receive queues. This condition typically occurs when the physical adapter enters a state from which it cannot independently recover, prompting the driver to initiate a software-level workaround to maintain service. While the driver is successfully monitoring and attempting to mitigate the stall, the repetitive nature of these logs confirms an underlying hardware or firmware-level synchronization issue that requires a driver and firmware alignment to permanently resolve.

Resolution

1. Upgrade the driver to the latest version based on HCL (https://compatibilityguide.broadcom.com)

2. Engage the hardware vendor of Intel network adapter for this issue if it persists.