Cisco virtual Port-channels technology is a pretty straight-forward technique: make two switches seem like one switch to a downstream switch, connected with a port-channel. The downstream switch has a single port-channel going upward, which makes the switch high available and eliminating the use of spanning-tree. It looks like this:

vPC in single N7k

Read more about vPC here. For the rest of this post, I’ll assume you know how to configure and use vPC.

In a lab where you have a single Nexus 7000 switch for testing and studying, you would divide it up into VDCs for ultimate usage. OTV, Fabricpath and the likes work out of the box, but if you configure vPC between two VDCs and use the management network (mgmt0) for the vPC keep-alive, you will be confronted with this message:

This has a pretty simple solution (besides buying a second Nexus 7k or using a dedicated keep-alive link). If you add a static route for the other VDCs management IP address towards the default gateway of your management network, communication will be possible. Example follows:

VDC1 Management IP:
VDC2 Management IP:
Management default gateway:

And on the second VDC:

After applying these static routes, your vPC adjacency should come up:

Share the wealth!