A hot standby node belongs to a clique of nodes and does not host virtual processes such as AMPs or Parsing Engines.
The standby node remains idle until a member of its clique crashes or fails. At this point, the standby node is activated to take over the failed node’s responsibilities.
The AMPs from the node that crashed are swiftly migrated to the hot standby node, resulting in negligible performance loss.
The intelligence of the design lies in the fact that once the failed node goes back online, it automatically takes over as the active standby node.