Failover causes a slave node to take over serving slots from the current master.
To initiate a failover, navigate to a slave node and select “Failover” from its node command menu.
The default failover mode performs a “safe” failover by coordinating with the current master and the rest of the cluster to ensure no writes are lost.
Force causes the slave to take over immediately without coordinating with the master and should be used when the master is not reachable.
Takeover causes the slave to take over immediately without coordinating with the master or the rest of the cluster. A new configuration epoch is generated unilaterally which may result in a collision.