This one really has me puzzled so I’m reaching out to the big brain group.
ProLiant DL560 Gen10, running ESXi 7 in a vSphere cluster. I have 3 HP Ethernet 10Gb 2-port 530SFP+ Adapters, with 10Gb connections.
What I’m seeing is that vmNIC3 is showing a disconnected status.
I’ve shuffled the fiber and SFP from a working port into this port and the problem remains with vmNIC3, proving that the problem is not the cabling or switch or SFP.
I’ve moved the NIC to a different PCI riser board and the problem remains with vmNIC3, proving that the problem is not with a bad riser board.
I’ve swapped in a different NIC and the problem remains with vmNIC3, proving that the problem is not a bad NIC.
Is it even possible that, in any way conceivable in and beyond the universe, something on the mainboard could cause only a single port on a NIC -a port in the middle of the range- to not work?
What am I overlooking?
Do I call Broadcom or HPE?
3 Spice ups
user35092
(user35092)
2
Are they both going to the same vSwitch? Make sure the NICs are NOT in standby mode.
1 Spice up
phildrew
(phildrew)
3
The NIC, since it’s HPE branded and came with the server, would be covered under HPE support for the server. However, from your description, this is not a hardware error. On the other hand - it can’t hurt to open a support case with them.
What do the switch status/logs show for that switchport? Is there a speed mismatch? What does iLO show for the NIC status? Is it disconnected there?
Check for NIC firmware and driver updates in VMware. I’ve seen some absolutely crazy network errors fixed by firmware and drivers before - review the release notes, don’t blindly update.
1 Spice up
The three 2-port NICs are identical models on identical firmware, so that seems unlikely to be a problem.
iLO shows that NIC status as “unknown” whereas all the others on the host show OK. i am a little curious why the OK one on this card shows neither IP4 nor IP6 addresses, while the other cards do.
A bit I don’t believe i covered is that these six ports are set to be 3 redundant connections. NIC2 and 5, NIC3 and 6, NIC4 and 7. in this picture you can see that 4 and 7 are associated:
On the iLO of a different host
I am going to have to look into this, see where/how it is set, perhaps this is a piece of the puzzle. Honestly, this is new territory for me, so a bit slow going.
BAM!!! 
The vmnic3 was administratively down.
esxcli network nic up -n vmnic3
This command brought it up and now it is working!
1 Spice up
phildrew
(phildrew)
8
Had three servers in a VMware cluster, all with the same NIC and driver/firmware. The servers were not all completely identical, but the networking was.
The issue only occurred with one host server.
