Our ESXI6.5 worked fine for several months.
The uplink to the physical switch is done by 3 physical NICs (vmnic3, vmnic4, vmnic5) in failover configuration.
Load balancing mode: explicit failover order
Network failure detection: Beacon Probing
Notify Switches: tried both Yes or No makes no difference
Failback: Yes
With the upgrade U1 installed we get every few minutes messages like:
Lost network connectivity on virtual switch "red". Physical NIC vmnic5 is down. Affected portgroups:"red".
The funny thing is, that the affected NIC changes from time to time. Any of 3 NICs show the Errors for a few hours before it changes to the next one.
Sometime 2 of them show the Errors alternating.
I don't think there is a hardware failure because any of the NICs show this behaviour and the actual connectivity is most of the time fine.
The statistics of the connected physical switch shows no Errors or Disconnects on any of the 3 ports who are connected to the physical NICs.
Any idea ?