Thank you, that's what I was hoping to hear.. I did some testing by making changes on a host with some running VM's, but the test only had a couple of VM's on it, our main production hosts have about 20+ VM's going at once. I really don't want to experiment with one of those hosts..
The question about "losing connectivity" was to address the scenario you mentioned; not doing any workaround, just reconfiguring the vMotion kernel port after they're done. So of a host tries to vmotion something and can't because the vmotion network isn't connecting to the target, it will just error and leave everything as is?
I'll probably take a combination of both approaches, set DRS to manual and don't try any vMotions while they're doing the work and then making the changes when they're done, but if it goes too long, go back and enable vmotion on the management kernel.