Cassandra failover vs other databases?

Cassandra offers controlled consistency like "write to 2 nodes and tell me it's done". Two "master" nodes and some slaves makes system good failover.

MongoDB offers replication pairs - simmilar failover force like cassandra? Is there any other databa开发者_如何学编程se with this form-box functionality?

Cassandra is a fully distributed system, so there is no need for explicit failover. If the machine you are sending requests to dies, you just reconnect to another (RRDNS, haproxy, any method is fine). Even losing an entire datacenter is handled by Cassandra without your app having to care.