Disaster Recovery - Run High Availability Across Data Centers

Based on DR Strategy 3, documented here: https://community.appian.com/success/w/guide/3329/high-availability-and-disaster-recovery-configurations#disaster_recovery, it sounds like it's possible for an HA environment to loose nodes and continue functioning, with some potential updates to configurations. 

• Has anyone tested this process?
• If you have 3 engine/data servers, if 1 node goes does, can 2 continue functioning? Likewise, if 2 nodes go down, can 1 function alone?
• What about search and app servers?