情形1：建一个topic:mytopic共3个分区，3个ReplicationFactor
bin/kafka-topics.sh --create --zookeeper 192.168.135.130:2181,192.168.135.129:2181,192.168.135.131:2181 --replication-factor 3 --partitions 3 --topic mytopic

Topic:mytopic PartitionCount:3 ReplicationFactor:3 Configs:
Topic: mytopic Partition: 0 Leader: 1 Replicas: 1,2,0 Isr: 1,2,0
Topic: mytopic Partition: 1 Leader: 2 Replicas: 2,0,1 Isr: 2,0,1
Topic: mytopic Partition: 2 Leader: 0 Replicas: 0,1,2 Isr: 0,1,2

如果挂掉分区0对应的1号节点，则可以正常发送消息

情形2：建一个topic:yourtopic共3个分区，1个ReplicationFactor
bin/kafka-topics.sh --create --zookeeper 192.168.135.130:2181,192.168.135.129:2181,192.168.135.131:2181 --replication-factor 1 --partitions 3 --topic yourtopic

Topic:yourtopic PartitionCount:3 ReplicationFactor:1 Configs:
Topic: yourtopic Partition: 0 Leader: 2 Replicas: 2 Isr: 2
Topic: yourtopic Partition: 1 Leader: 0 Replicas: 0 Isr: 0
Topic: yourtopic Partition: 2 Leader: 1 Replicas: 1 Isr: 1

如果挂掉分区0对应的2号节点，则不能正常发送消息，报错：Error: partition 0 does not have a leader. Skip getting offsets

ITPub博客

验证kafka集群高可用

最新文章