VMware vSAN 8.x
Intel(R) Ethernet Controller X710 for 10GbE SFP+
The issue is caused by transmit (TX) ring buffer exhaustion on the physical network adapter (Intel X710). This occurs when bursty vSAN traffic exceeds the driver's currently allocated buffer capacity.
When the TX ring buffer is saturated, the driver cannot accept new packets from the VMkernel, causing the txBusy and queueFull counters to increment. The restartQueue value indicates the driver had to reset the transmission logic to recover from this state.
vmnic3 stats :
txq0: totalPkts=50889064629 totalBytes=181267271786288 restartQueue=11 txBusy=10 queueFull=10 pktDropped=0
Run the following command on the affected ESXi host to identify the error type for the specific vmnic (e.g vmnic3)
# esxcli network nic stats get -n vmnic3
Review the output for the following indicators:
restartQueue > 0txBusy > 0queueFull > 0# esxcli network nic ring current get -n vmnic3
# esxcli network nic ring preset get -n vmnic3
Example findings on Intel X710
Current RX:1024, TX:1024Max Supported: Max RX:4096, Max TX: 4096If the error counters continue to increment during peak workloads, expand the ring buffer sizes to the maximum supported by the hardware driver to better absorb traffic bursts:
Review this command before running it.
# esxcli network nic ring current set -n vmnic3 -r 4096 -t 4096
Note: Increasing ring buffers uses more system memory. Ensure the host has adequate resources before applying.
If the alarm persists after buffer adjustment, engage the hardware vendor and physical networking team to inspect: