I'm attempting to test MTU changes with various functions (SCP/vMotion/etc...) so I wanted to create a second management interface that wouldn't disable the other one since I don't have console access to the ESXi host.
I created a new VMkernel nic and gave it a static IP (10.170.2.93/24); I added Managment/vMotion options to the interface. Then I added it to a new vSwitch. Both the new VMkernel nic and vSwitch have MTU set to 4088 (inching to 9000). The vSwitch has two uplinks associated to it. There is another existing vSwitch on the same server with two other uplinks where the existing management network resides (different subnets).
Directly patched to the server, I can ping the above IP address, but when I run a port scan, no ports are open. Did I go in the wrong direction here with what I was trying to accomplish? Or is it something more simple like a daemon restart that I didn't do?
With what I see from your explanation , Did you happen to check from esxi if those ports are really closed which you did see as closed from port scan ?
Did you validate for which purpose those particular ports are used for ?
If it is closed from esxi , then did you happen to try to open those ports ?