When static load sharing is used for the MLAG
ports, and if there is a single link connecting the server node and the MLAG peer switches,
the port does not need to be configured as a load shared port on the MLAG peer switches.
Configuring LACP on MLAG ports can reduce the number of load shared ports that can be
configured in the system.
LACP configuration for MLAG ports (system priority, LACP timeout,
activity mode, etc.) should be the same on all the MLAG peer switches.
We recommend that the server node has a lower System Aggregation
Priority than the MLAG peers so that the server node chooses which of the ports can be
aggregated. As an example, there are a maximum of 8 ports that can be aggregated together,
and there are eight links from Peer1 to the Server, and another eight links from Peer2 to
the server node. When the server node has a lower System Aggregation Priority, it can
choose which of the total available links can be aggregated together.
If the Port Aggregation Priority is not configured for the load
shared member ports, there is a chance that only the links from server node to one of the
MLAG peer are aggregated together (based on the port numbers). In this instance, the links
from the server node to the other MLAG peer are unused. To avoid this, you can configure
the Port Aggregation Priority on the server node so that the number of active links to the
MLAG peers is balanced.
You must configure load sharing groups on all the MLAG ports even if
they contain just one port.
It is recommended to configure static LACP MAC on both MLAG peers to avoid traffic
interruption during failure scenarios.