Split Brain and Network Partitions

1. A network failure splits a distributed cluster into two isolated groups. Each group thinks the other is down and continues accepting writes independently. What problem has occurred?

2. To prevent split-brain, the cluster requires that a primary can only be active if it can communicate with a majority of nodes. What mechanism is this?

3. To protect against a stale primary accepting writes after a partition, the system assigns each primary a time-bounded lease. The primary must renew the lease by communicating with followers. If the lease expires, it must stop accepting writes. What is this called?

4. The CAP theorem states that a distributed system can only guarantee two of three properties: Consistency, Availability, and Partition Tolerance. If the system prioritises Consistency and Partition Tolerance (CP), what must it sacrifice during a partition?

5. A primary database replica detects a partition but cannot determine if it has been replaced by a new leader. To prevent both nodes from writing, the system sends a STONITH (Shoot The Other Node In The Head) command. What is this technique called?