History

Alyssa Huang 7339d65b27 KAFKA-19354: KRaft observer should send fetch to best node (#19854 ) Observers may get stuck fetching from bootstrap servers even on discovery of a leader from a fetch response. Observers currently fetch from bootstrap servers once their fetch timeout expires, and even on rediscovery of the leader (e.g. observer receives fetch response indicating location of leader), observers will not attempt to resume fetching from the leader. This change simplifies observer polling logic to just rely on maybeSendFetchToBestNode which takes care of the logic of selecting either the leader or bootstrap servers as the destination based on timeouts and backoffs. Reviewers: José Armando García Sancio <jsancio@apache.org>		2025-07-28 13:03:14 -04:00
..
bin	KAFKA-18388 test-kraft-server-start.sh should use log4j2.yaml (#18370 )	2025-01-07 04:20:06 +08:00
config	KAFKA-18388 test-kraft-server-start.sh should use log4j2.yaml (#18370 )	2025-01-07 04:20:06 +08:00
src	KAFKA-19354: KRaft observer should send fetch to best node (#19854 )	2025-07-28 13:03:14 -04:00
.gitignore	KAFKA-13429: ignore bin on new modules (#11415 )	2021-11-10 14:36:24 -06:00
README.md	KAFKA-17728 Add missing config `replica-directory-id` to raft README (#17518 )	2024-10-17 11:32:27 +08:00

README.md

KRaft (Kafka Raft)

KRaft (Kafka Raft) is a protocol based on the Raft Consensus Protocol tailored for Apache Kafka.

This is used by Apache Kafka in the KRaft (Kafka Raft Metadata) mode. We also have a standalone test server which can be used for performance testing. We describe the details to set this up below.

Run Single Quorum

bin/test-kraft-server-start.sh --config config/kraft.properties --replica-directory-id b8tRS7h4TJ2Vt43Dp85v2A

Run Multi Node Quorum

Create 3 separate KRaft quorum properties as the following:

cat << EOF >> config/kraft-quorum-1.properties

node.id=1
listeners=PLAINTEXT://localhost:9092
controller.listener.names=PLAINTEXT
controller.quorum.voters=1@localhost:9092,2@localhost:9093,3@localhost:9094
log.dirs=/tmp/kraft-logs-1
EOF

cat << EOF >> config/kraft-quorum-2.properties

node.id=2
listeners=PLAINTEXT://localhost:9093
controller.listener.names=PLAINTEXT
controller.quorum.voters=1@localhost:9092,2@localhost:9093,3@localhost:9094
log.dirs=/tmp/kraft-logs-2
EOF

cat << EOF >> config/kraft-quorum-3.properties

node.id=3
listeners=PLAINTEXT://localhost:9094
controller.listener.names=PLAINTEXT
controller.quorum.voters=1@localhost:9092,2@localhost:9093,3@localhost:9094
log.dirs=/tmp/kraft-logs-3
EOF

Open up 3 separate terminals, and run individual commands:

bin/test-kraft-server-start.sh --config config/kraft-quorum-1.properties --replica-directory-id b8tRS7h4TJ2Vt43Dp85v2A
bin/test-kraft-server-start.sh --config config/kraft-quorum-2.properties --replica-directory-id Nkij_D9XRiYKNb41SiJo7Q
bin/test-kraft-server-start.sh --config config/kraft-quorum-3.properties --replica-directory-id 4-e97nI7eHPYKfEDtW8rtQ

Once a leader is elected, it will begin writing to an internal __raft_performance_test topic with a steady workload of random data. You can control the workload using the --throughput and --record-size arguments passed to test-kraft-server-start.sh.