ESXi host looses network connectivity post updating firmware of UCS/VIC due to an incorrect UCS vNIC template configuration

Products

• VMware vSphere ESXi

Issue/Introduction

• Test Management Network fails in DCUI
• Host does not respond to ping on management vmkernel interface
• Uplink interfaces (for example, vmnic0/vmnic1) show TX packets only with no RX packets
• Repeated NIC resets observed in /var/run/log/vmkernel.log

  WARNING: Uplink: Queue 0 of device vmnic# stuck, resetting the device
  nenic: enic_uplink_reset: Resetting
  nenic: enic_dev_log_qerror_capable: CMD_QUEUE_ERROR not supported

• Watchdog timeout events for affected vmnic's observed in /var/run/log/vobd.log

  vob.net.uplink.watchdog.timeout: Watchdog timeout occurred for uplink vmnic#

Environment

VMware vSphere ESXi 

Resolution

In the UCS Manager vNIC template: Uncheck the no support failover option (This option can cause unwanted vNIC remapping and unstable vmnic ordering.)

After making changes, reboot the blade to apply the updated profile.

Enabling CDN ensures predictable vmnic enumeration during boot.

Cisco documentation describes how to enable CDN in BIOS policy: Cisco UCS Manager Server Management Guide