Accepted Solutions
So phyical NICs don't have IP addresses assigned to them in ESXi. You tell ESXi what phy nics should be plugged into which virtual switch and you assign an ip address to the network management agent that's also plugged into that same virtual switch, but you never assign IP addresses to physical nics. The only purpose of a phyical nic in an ESXi install is so they can plug into virtual switches you create inside of your ESXi install. Then it's the other VMs and VMKernel ports that have their own IP addresses. Take this image below I found on the internet as an example. Those physical nics (vmnic0 and vmnic1) don't have any ip addresses, there's no DNS A record pointing to those, etc. They are there to plug physical switches into your virtual switches (they are network devices; that's the clearest way to think about them). The grey bar below is a virtual switch and the green dots on the virtual switch are ports where something is plugged into it. Those VMs (Linux, Windows, whatever) is what has ipaddresses. The special devices plugged into the virtual switch are called VMKernel Ports and Service Console Ports (in ESXi 4.x they are all just called VMKernel Ports). These have IP addresses and those are the closest things to what you are describing in your post.
If you get the graphic though, you'll notice that the virtual switch "bonds" those two phy nics together based on how you tell it to bond them. Unless you want to do a lot of research or you have a complex, exceptional case, the settings for bonding those two nics (in the "NIC Teaming" tab) probably ought to be "Route based on the originating virtual port ID" and "Link status only". Beacon probing should only be used in cases when Link status only mode don't work (usually it does just fine). Furthermore I also set both NICs as "Active" and say yes to "Notify switches" and "Fallback".
So phyical NICs don't have IP addresses assigned to them in ESXi. You tell ESXi what phy nics should be plugged into which virtual switch and you assign an ip address to the network management agent that's also plugged into that same virtual switch, but you never assign IP addresses to physical nics. The only purpose of a phyical nic in an ESXi install is so they can plug into virtual switches you create inside of your ESXi install. Then it's the other VMs and VMKernel ports that have their own IP addresses. Take this image below I found on the internet as an example. Those physical nics (vmnic0 and vmnic1) don't have any ip addresses, there's no DNS A record pointing to those, etc. They are there to plug physical switches into your virtual switches (they are network devices; that's the clearest way to think about them). The grey bar below is a virtual switch and the green dots on the virtual switch are ports where something is plugged into it. Those VMs (Linux, Windows, whatever) is what has ipaddresses. The special devices plugged into the virtual switch are called VMKernel Ports and Service Console Ports (in ESXi 4.x they are all just called VMKernel Ports). These have IP addresses and those are the closest things to what you are describing in your post.
If you get the graphic though, you'll notice that the virtual switch "bonds" those two phy nics together based on how you tell it to bond them. Unless you want to do a lot of research or you have a complex, exceptional case, the settings for bonding those two nics (in the "NIC Teaming" tab) probably ought to be "Route based on the originating virtual port ID" and "Link status only". Beacon probing should only be used in cases when Link status only mode don't work (usually it does just fine). Furthermore I also set both NICs as "Active" and say yes to "Notify switches" and "Fallback".
