In a two node stretch metro cluster configuration, what is the expected failover behavior during a site outage (e.g. power outage to one controller and its stack of disk shelves)?
Is it automatic, the same as you expect from an HA-pair, or is it manually initiated (e.g. cf forcetakeover, etc.)?
An Automatic MetroCluster Site Failover Solution
Nicholas Lee Fagan
that's not true !
as soon either the partner node nor the shelves will be seen, no automatic TO will happen !