Hi,
Yesterday I've encountered and very odd issue. I have 2 ESX hosts with 5 Gbit NICs (vmnic0 .. vmnic4). I've configured dvSwitch with 5 uplinks (uplink1 - uplink5) mapped to (vmnic0 - vmnic4). I've created various portgroups on dvSwitch (configured each portgroup with only one uplink!) and some virtual adapters on the hosts which use those portgroups. All works fine.
Than I added 2 more hosts to my cluster, but those are different than the first two, vmnics are differently numbered (but I have the same UTP cabling for the sake of transparent cabling in the rack). Because of this I had to change the uplink-vmnic mapping on those 2 additional hosts.
That said, I have:
first 2 hosts:
uplink1 = vmnic0 (on-board dual-port)
uplink2 = vmnic1 (on-board dual-port)
uplink3 = vmnic2 (Intel PCI-e dual-port)
uplink4 = vmnic3 (Intel PCI-e dual-port)
uplink5 = vmnic4 (Broadcom PCI-e Single-Port)
second 2 hosts:
uplink1 = vmnic1 (Intel PCI-X dual-port)
uplink2 = vmnic4 (on-board dual-port)
uplink3 = vmnic0 (Intel PCI-X dual-port)
uplink4 = vmnic2 (Intel PCI-e Single-Port)
uplink5 = vmnic3 (on-board dual-port)
Let me just point out that the cabling and mapping is 100% correct, I've double-checked many times.
Now, because portgroups are linked to uplinks and NOT vmnics I thought this should be all transparent and working configuration, however for me this is not the case.
The odd issues are:
1. When I add one of the second 2 hosts to the dvswitch (with proper mapping and migration of the default standard switch with management console virtual adapter), vCenter looses connection to the hosts (I can't even ping the host). I go to the hosts physical console, do the "Restart management network" and the host is connected again.
2. When I reboot the server, It's management network is not working (I can't even ping the host). I go to the hosts physical console, "Restart management network" and the host is connected again.
3. The "Hardware status" of the host is not working (I don't know how this is related to this issue, but so it is).
If I then go and remove host from dvSwitch and "Restore standard switch", everything is working OK. If I reboot the server, connection is alive after reboot and the "Hardware status" of the host is working fine. If I go again through the scenario of adding host to dvSwitch, the same problems occur again.
I have NO problems with the first 2 hosts, everyting is working fine, even after reboot, hardware status is OK, etc.
Are there any special settings for the dvSwitch when using different mapping uplink-vmnic on different hosts? Thanks for help!