KRaft — Kafka Without ZooKeeper
How KRaft replaced ZooKeeper, enabling Kafka to manage its own metadata via Raft consensus.
0 of 5 lessons complete
- 1Why ZooKeeper Had to GoBefore KRaft, Kafka depended on Apache ZooKeeper for cluster metadata such as broker registration, controller election, …
- 2Raft Consensus & the Controller QuorumKRaft uses the Raft consensus algorithm. A small set of servers, typically 3 or 5, is configured as controllers and form…
- 3The Metadata LogIn KRaft the cluster metadata lives in an internal single-partition topic, __cluster_metadata, whose replication is mana…
- 4Controller Epochs & FencingEach leader election in the controller quorum increments the leader epoch, a monotonically increasing number; the monito…
- 5KRaft vs ZooKeeper: PerformanceKIP-500 states what KRaft is meant to improve. It aims to manage metadata in a more scalable and robust way, with suppor…
Knowledge check
Q1.What is the role of the leader (quorum) epoch in the KRaft controller quorum?
Q2.How many controller nodes are needed to tolerate 1 failure?
Q3.Where is cluster metadata stored in KRaft mode?