Multiple ChoiceHard 45s
Fencing tokens for stale leaders
A stalled old leader resumes and writes to shared storage after a new leader was elected. What mechanism prevents the stale leader's writes from corrupting state?
Options
- 1A longer heartbeat interval on the leader, so followers wait longer before declaring it dead and can reject its late writes
- 2Read replicas that serve stale data, so a partitioned leader keeps writing to them without any storage-level rejection
- 3Monotonically increasing fencing tokens rejected by storage if not greater than the last seen token
- 4Two-phase commit across all replicas, so storage blocks the stale leader's writes at prepare and aborts them before commit
Think you can answer under pressure?
Get matched and answer live questions in 3-minute 1v1 battles. Elo ranking, badge tiers and bragging rights on the line.