Hey, all. So, we've got some OpenStack controllers that I'm
virtualizing, and it's going really well... until the most recent host I
was dooing, the fifth of eight. Everything looked like it largely
should, except that there was zero traffic on the VLANs on the VM.
Sniffing physical ports showed it hitting the machine, but nothing made
it into VLAN/namespace land on the VMs. After banging my head against
the wall a lot, I realized that, for some reason, the interface team
(which is hot/cold) was using the secondary interface, not the primary.
I turned that interface down, the primary came up, and everything
started working great. (It only became apparent that something like
that might be the case when I compared the broken machine's output from
"brctl listmacs br0", vs. that of a working machine.)
Any idea what's _really_ not working, deep down on the stack? Is there
a way around it?
Thanks!
-Ken