AutoBGP LAG Active/Standby Mode

For each Link Aggregation Control Protocol (LACP)-attached port, information on the port is distributed through the auto-peering network using a proprietary Border Gateway Protocol (BGP)-based mechanism. It uses the private AFI/SAFI to propagate the Ethernet segment ID and remote router ID as next hop.

This information is used, with a deterministic mechanism (lowest router ID), by a switch to place its port in standby mode. Standby mode means "Oper Status Down," and link down. If the active port goes down, the switch in the standby mode detects this using the same BGP mechanism, and makes its port active. Packets are reflected until BGP convergence of Ethernet segment ID to participating nodes.

For each attached client, only one attached port per switch is allowed. By default, after enabling auto-peering, each switch automatically supports creation of LAGs on all edge (client) ports, and prevents creation of LAGs on all auto-peering ports.

After BGP auto-peering is enabled on a switch (create auto-peering bgp routerid ipaddress AS-number asNumber ), the equivalent of the following configuration is enacted for every Ethernet port (X) on the switch:
enable sharing X grouping X lacp 
configure sharing X lacp fallback enable 
configure sharing X lacp fallback timeout 1

This allows a LAG to come up on LAG-capable attached devices, and to have the port become a forwarding bridge port in one second if attached to a non-LAG capable (or LAG is disabled) device. If the attached device later becomes able to support LAG, LAG automatically comes up. In the same manner, as manually executing the commands, this produces a persistent state, and the commands appear in the configuration. This configuration persists after reboots.

If auto-peering is disabled on a switch, the equivalent of the following command is automatically executed:
disable sharing X