Intel NIC occasionally experiences link down issues after upgrading ixgben driver to version 1.19.3.0 or 1.20.2.0

Products

• VMware vSphere ESXi

Issue/Introduction

• After upgrading Intel’s ixgben async driver to version 1.19.3.0 or 1.20.2.0, the Intel NIC occasionally experiences link down issues.

• The physical network has remained unchanged.

• This may be seen primarily or solely as occurring in conjunction with VM vMotions.

• The following error messages may have been recorded in the ESXi /var/run/log/vobd.log logs:

  vobd[2098148]:  [netCorrelator] 41456737us: [vob.net.vmnic.linkstate.down] vmnic vmnicX linkstate down

"Lost uplink redundancy on virtual switch <vSwitch_name>. Physical NIC <vmnic_ID> is down. Affected portgroups: <list_of_portgroups>".
• The following error messages have been recorded in the ESXi /var/run/log/vmkernel.log logs:
  
  vmkwarning: cpu#:#######)WARNING: ixgben: ixgben_CheckTxHang:2181: 0000:18:00.0: TX hang detected on TX queue
vmkernel: cpu#:#######)ixgben: ixgben_ServiceTimer:3609: 0000:18:00.0: TX hang detected: reset device
vmkernel: cpu#:#######)ixgben: ixgben_CheckLink:405: 0000:18:00.0: Link is down for device

Environment

VMware vSphere ESXi 8.0

Cause

A known defect in Intel ixgben driver versions 1.19.3.0 and 1.20.2.0 causes false transmit (TX) hangs and subsequent hardware reset commands, which drop the physical link.

Resolution

1. Navigate to the Broadcom Support Portal.
2. Download the ixgben async driver version 1.22.1.0 or newer.
   1. See Download Broadcom products and software for detailed download instructions.
3. Install the updated driver on all affected ESXi hosts.
   1. See Download and install async drivers in VMware ESXi
4. If the issue persists, engage your hardware vendor to verify and update NIC firmware to a compatible version.

Additional Information

Intel NIC で ixgben ドライバをバージョン 1.19.3.0 もしくは 1.20.2.0 にアップグレードした後、時折リンクダウンの問題が発生する