On Wed, Jun 12, 2019 at 02:45:54PM +0000, Alexander Vainshtein wrote: > Albrecht, Reshad and all, > I concur with Reshad - it will work (RFC 7130 never says anything about the > number of links in the LAG).
A somewhat perverse use of the feature. Cute. :-) > Such a session would still use encapsulation in IP/UDP with the UDP > destination port set to 6784 and the Destination IP address " that is > configured on the peer system and can be reached via the LAG interface ". > I.e., some lightweight IP functionality would be still required. I'd expect there to be two interesting headaches: 1. In some implementations, you're shutting down "the lag" rather than the link. You'll effectively stop forwarding traffic, I think, but you may not get the expected ifDown event on the link, only on the LAG-of-1 virtual interface. Clearly this is an implementation detail, but such things are what make LAGs painful in IP protocols land. 2. The bootstrapping provisions require a bit more than just the light-weight IP stack. My employer's developers liked to say that they had to operate at "layer 2.5" in order to get stuff to work before IP was actually up. -- Jeff
