kafka-reassign-partitions --generate for topic __commit_offsets 给了我奇怪的结果：分区副本只在一个代理上

Question

kafka-reassign-partitions --generate for topic __commit_offsets gives me strange result: partition replica only on one broker anyway, but expected 3 replicas for each partition.

I do the following:

test@kafka-1:~/Kafka-Docker$ sudo docker exec -it broker bash
[appuser@broker ~]$ echo '{"version":1, "topics":[{"topic":"__consumer_offsets"}]}' > topics-to-move.json
[appuser@broker ~]$ kafka-reassign-partitions --zookeeper $KAFKA_ZOOKEEPER_CONNECT --topics-to-move-json-file topics-to-move.json --broker-list "1,2,3" --generate
Warning: --zookeeper is deprecated, and will be removed in a future version of Kafka.
Current partition replica assignment
{"version":1,"partitions":[{"topic":"__consumer_offsets","partition":0,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":1,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":2,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":3,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":4,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":5,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":6,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":7,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":8,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":9,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":10,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":11,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":12,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":13,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":14,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":15,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":16,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":17,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":18,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":19,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":20,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":21,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":22,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":23,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":24,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":25,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":26,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":27,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":28,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":29,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":30,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":31,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":32,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":33,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":34,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":35,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":36,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":37,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":38,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":39,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":40,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":41,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":42,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":43,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":44,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":45,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":46,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":47,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":48,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":49,"replicas":[1],"log_dirs":["any"]}]}

Proposed partition reassignment configuration
{"version":1,"partitions":[{"topic":"__consumer_offsets","partition":0,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":1,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":2,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":3,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":4,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":5,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":6,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":7,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":8,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":9,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":10,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":11,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":12,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":13,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":14,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":15,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":16,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":17,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":18,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":19,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":20,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":21,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":22,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":23,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":24,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":25,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":26,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":27,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":28,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":29,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":30,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":31,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":32,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":33,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":34,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":35,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":36,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":37,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":38,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":39,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":40,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":41,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":42,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":43,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":44,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":45,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":46,"replicas":[2],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":47,"replicas":[3],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":48,"replicas":[1],"log_dirs":["any"]},{"topic":"__consumer_offsets","partition":49,"replicas":[2],"log_dirs":["any"]}]}
[appuser@broker ~]$

I see that zookeepers option is deprecated so I also tried this (with the same result):

kafka-reassign-partitions --bootstrap-server kafka-1:9092  --topics-to-move-json-file topics-to-move.json --broker-list "1,2,3" --generate

PROBLEM:

As you may see, "Current partition replica assignment" has "replicas: [1]" only and "Proposed partition reassignment configuration" has "replicas: [1]" or "replicas: [2]" or "replicas: [3]" for each partition.

But for "Proposed partition reassignment configuration" it expected to be something like "replicas: [1,2,3]" or "replicas: [2,3,1]" and so on for redudancy of __commit_offsets across 3 brokers.

Why only one replica for each partiotion is bound to one broker only? Why 3 replicas for each partition are not spreaded across 3 brokers?

Example of good topic (but it has only one partition) replicated across all 3 brokers, it looks this way:

{"version":1,"partitions":[{"topic":"test_topic","partition":0,"replicas":[1,2,3],"log_dirs":["any","any","any"]}]}

My setup: 3 VMs 192.168.1.11, .12 and.13 each has zookeeper and broker, zookeepers connected to quorum ok - I see only one leader through its REST API (for example http://192.168.1.11:4888/commands/stats same for .12 and .13 IPs) and that's good. Also tested switch off one of zookeepers and new leader elected in that case - works perfect.

Example of docker-compose.yml for 192.168.1.11 VM is (other similar):

version: '3.7'

x-zoo: &zoo "kafka-1:2888:3888;kafka-2:2888:3888;kafka-3:2888:3888"
x-kafkaZookeepers: &kafkaZookeepers "kafka-1:2181,kafka-2:2181,kafka-3:2181"
x-kafkaBrokers: &kafkaBrokers "kafka-1:9092,kafka-2:9092,kafka-3:9092"

services:
  zookeeper:
    image: confluentinc/cp-zookeeper:6.2.1
    hostname: zookeeper
    container_name: zookeeper
    ports:
      - "2181:2181"
      - "2888:2888"
      - "3888:3888"
      - "4888:8080"
    extra_hosts:
      - "kafka-1:192.168.1.11"
      - "kafka-2:192.168.1.12"
      - "kafka-3:192.168.1.13"
    environment:
      ZOOKEEPER_SERVER_ID: 1
      ZOOKEEPER_QUORUM_LISTEN_ON_ALL_IPS: 'true'
      ZOOKEEPER_CLIENT_PORT: 2181
      ZOOKEEPER_PEER_PORT: 2888
      ZOOKEEPER_LEADER_PORT: 3888
      ZOOKEEPER_TICK_TIME: 2000
      ZOOKEEPER_INIT_LIMIT: 5
      ZOOKEEPER_SYNC_LIMIT: 2
      ZOOKEEPER_SERVERS: *zoo
    volumes:
      - ./kafka-data/zookeeper:/var/lib/zookeeper/data
      - ./kafka-data/zookeeper-logs:/var/lib/zookeeper/log
    networks:
      - mynet

  broker:
    image: confluentinc/cp-kafka:6.2.1
    hostname: broker
    container_name: broker
    depends_on:
      - zookeeper
    ports:
      - "9092:9092"
    extra_hosts:
      - "kafka-1:192.168.1.11"
      - "kafka-2:192.168.1.12"
      - "kafka-3:192.168.1.13"
    environment:
      KAFKA_BROKER_ID: 1
      KAFKA_LISTENERS: OUTSIDE://0.0.0.0:9092
      KAFKA_LISTENER_SECURITY_PROTOCOL_MAP: OUTSIDE:PLAINTEXT
      KAFKA_ADVERTISED_LISTENERS: OUTSIDE://192.168.1.11:9092
      KAFKA_INTER_BROKER_LISTENER_NAME: OUTSIDE
      KAFKA_ZOOKEEPER_CONNECT: *kafkaZookeepers
      KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: 3
      KAFKA_DEFAULT_REPLICATION_FACTOR: 3
      KAFKA_LOG_MESSAGE_TIMESTAMP_TYPE: 'LogAppendTime'
      KAFKA_GROUP_INITIAL_REBALANCE_DELAY_MS: 0
      KAFKA_CONNECTIONS_MAX_IDLE_MS: 31536000000 # 1 year
    volumes:
      - ./kafka-data/kafka:/var/lib/kafka/data
    networks:
      - mynet

networks:
  mynet:
    driver: bridge

It was set up first with:

1. KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: 1
2. KAFKA_DEFAULT_REPLICATION_FACTOR: 1

Then +2 other setups on .12 and .13 VMs added (of course with different ZOOKEEPER_SERVER_ID and KAFKA_BROKER_ID ) and on each KAFKA_DEFAULT_REPLICATION_FACTOR and KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR changed from 1 to 3.

Then new topics created OK with replication factor 3, also I was able easily change replication factor to 3 for my existing non-system topics, but except topic __commit_offsets of cource - it need to be reassigned using kafka-reassign-partitions command line tool.

Answer 1

The offsets topic only gets created on a brand new cluster. It will not retroactively get updated if you modify offsets.topic.replication.factor .

Similarly, any topic created before you increased default.replication.factor will also only have one replica.

Reassign partition tool doesn't take into account your server properties. You either need a different tool or manually place the replicas yourself. - How to change the number of replicas of a Kafka topic?