It's automatic. There is nothing to do under normal circumstances. Gateways/endpoints do not work like IPSI's. Nothing actually changes in the ESS, to make it "active". It just sits there waiting for something to try to connect. If a gateway or endpoint cannot reach procr, it consults it's mgc list and "finds" an ESS or LSP. If you have a particular site that you manually want to be controlled by the ESS even though the network is up (like before a maintenance window on the CORE) you would "disable nr X" in the CORE. Those endpoints and gateways will then be rejected by the CORE and will go find their appropriate ESS/LSP based on their mgc configuration. "Enable NR X" to allow the site back to the CORE. That's the closest equivalent to an ipsi-takeover.