MLAG needs to be able to determine the difference between two types of failures that appear very similar:
Peer switch fails
Peer-link has failed
MLAG uses the management network communication to determine if the other switch is still working when the peer-link has failed. If the other switch is still responding when the peerlink fails, then the secondary switch will take down all mlag interfaces until the peerlink is restored or until the primary switch fails.
If you do not have the VIP configured on the management network, then the switches will not properly handle peerlink failures and will go into the splitbrain state which relies on STP and LACP system MAC to handle failovers. If LACP on the clients does not notice that there is now two system MACs on different ports in the bond, or if the clients are using static port-channels instead of LACP, then a peerlink failure would result in an outage. If STP is not properly configured for this failure scenario, then an outage could happen as it is not a usual failure scenario where the management network and peerlink fail at the same time, and many networks are not designed to handle this double-failure splitbrain case using STP.
One more question, what is the VLAN requirement for the IPL port-channel?
This port-channel must be configured as “access”, correct? on any VLAN?
Can I use the same VLAN that I configured for the IPL VLAN interface (e.g., VLAN 4000) ?