kafka集群双活-数据迁移

一、集群搭建（kafka使用自带的zookeeper）

前提：必须要有java环境

1、下载地址：

2、安装目录 /app/kafka

 

建立数据和日志存储目录

mkdir -p /app/kafka/data/zookeeper/

mkdir -p /app/kafka/data/kafka/

mkdir -p /app/kafka/log/zookeeper/

mkdir -p /app/kafka/log/kafka/

 

cd /app/kafka/kafka-3.6.1-src

 

vim config/zookeeper.properties

 

配置如下：

dataDir=/app/kafka/data/zookeeper
dataLogDir=/app/kafka/log/zookeeper
# the port at which the clients will connect
clientPort=2181
# disable the per-ip limit on the number of connections since this is a non-production config
maxClientCnxns=100
# Disable the adminserver by default to avoid port conflicts.
# Set the port to something non-conflicting if choosing to enable this
admin.enableServer=false
# admin.serverPort=8080

tickTime=2000
initLimit=10
syncLimit=5

server.1=192.168.43.145:12888:13888
server.2=192.168.43.146:12888:13888
server.3=192.168.43.121:12888:13888

 

vim config/server.properties

 

broker.id=1
listeners=PLAINTEXT://192.168.43.145:9092 #另外两个节点修改成对应的ip
log.dirs=/app/kafka/log/kafka
zookeeper.connect=192.168.43.145:2181,192.168.43.146:2181,192.168.43.121:2181

 

 

所有节点启动服务

./zookeeper-server-start.sh -daemon ../config/zookeeper.properties

./kafka-server-start.sh -daemon ../config/server.properties

 

查看是否有报错，如果有报错，把防火墙关闭后，再启动

systemctl stop firewalld

systemctl start firewalld

systemctl status firewalld

 

验证：

sh ./zookeeper-shell.sh 192.168.43.121:2181

ls /brokers/ids

看到

【1，2，3】

说明集群构建成功

 

 

二、数据迁移和同步（mirror-maker mm2）

vim ../config/connect-mirror-maker.properties

 

clusters = A, B     #集群名称，代号

# connection information for each cluster
# This is a comma separated host:port pairs for each cluster
# for e.g. "A_host1:9092, A_host2:9092, A_host3:9092"
A.bootstrap.servers = A_host1:9092, A_host2:9092, A_host3:9092 #源集群
B.bootstrap.servers = B_host1:9092, B_host2:9092, B_host3:9092 #目标集群

# enable and configure individual replication flows
A->B.enabled = true  #同步方向 A集群->B集群

# regex which defines which topics gets replicated. For eg "foo-.*"
A->B.topics = .*    #同步主题过滤

#B->A.enabled = true  #建议注释掉 双向实时同步可能会导致cup,内存爆掉
#B->A.topics = .*     #建议注释掉 双向实时同步可能会导致cup,内存爆掉

#同步协议，不加默认同步的主题会带上 集群前缀例如 A.topic,B.topic
replication.policy.class=org.apache.kafka.connect.mirror.IdentityReplicationPolicy

# Setting replication factor of newly created remote topics

#两个集群保持一致
replication.factor=1

启动命令：

./connect-mirror-maker.sh -daemon ../config/connect-mirror-maker.propertie

 

 

验证：

查看主题列表：

./kafka-topics.sh --bootstrap-server 192.168.43.145:9092 --list

创建主题test

./kafka-topics.sh --bootstrap-server 192.168.43.145:9092 --create --topic test

 

 

#最大位移

./kafka-run-class.sh kafka.tools.GetOffsetShell --broker-list 192.168.43.146:9092 --topic test --time -1

 

#最早位移

./kafka-run-class.sh kafka.tools.GetOffsetShell --broker-list 192.168.43.146:9092 --topic test --time -2