首先下载Maxwell,Maxwell下载地址:https://github.com/zendesk/maxwell/releases/tag/v1.28.2,现在版本1.30.0版本之上需要jdk11以上,建议下载1.30.0版本以下版本。下载完成后按照如下步骤进行配置,同步MySQL数据到Kafka,前提是MySQL需要开启Binlog日志,可以参考Canal章节设置。Maxwell不支持高可用搭建,但是支持断点还原,可以在执行失败时重新启动继续上次位置读取数据。
将下载好的安装包上传到node3并解压[root@node3 ~]# cd /opt/apps [root@node3 software]# tar -zxvf ./maxwell-1.28.2.tar.gz -C /opt在MySQL中创建Maxwell的用户及赋权
Maxwell同步mysql数据到Kafka中需要将读取的binlog位置文件及位置信息等数据存入MySQL,所以这里创建maxwell数据库,及给maxwell用户赋权访问其他所有数据库。
mysql> CREATE database maxwell; mysql> CREATE USER 'maxwell'@'%' IDENTIFIED BY 'maxwell'; mysql> GRANT ALL ON maxwell.* TO 'maxwell'@'%'; mysql> GRANT SELECt, REPLICATION CLIENT, REPLICATION SLAVE ON *.* TO 'maxwell'@'%'; mysql> flush privileges;修改配置“config.properties”文件
进入“/opt/maxwell-1.28.2”,修改“config.properties.example”为“config.properties”并配置:
producer=kafka kafka.bootstrap.servers=node1:9092,node2:9092,node3:9092 kafka_topic=maxwell_topic #设置根据表将binlog写入Kafka不同分区,还可指定:[database, table, primary_key, transaction_id, thread_id, column] producer_partition_by=table #mysql 节点 host=node1 #连接mysql用户名和密码 user=maxwell password=maxwell
注意:以上参数也可以在后期启动maxwell时指定参数方式来设置。
启动zookeeper及Kafka,并监控Kafka maxwell_topic[root@node3 bin]# ./kafka-console-consumer.sh --bootstrap-server node1:9092,node2:9092,node3:9092 --topic maxwell_topic启动Maxwell
[root@node4 bin]# maxwell --config ../config.properties
注意以上启动也可以编写脚本:
#startMaxwell.sh 脚本内容: /opt/maxwell-1.28.2/bin/maxwell --config /opt/maxwell-1.28.2/config.properties > ./log.txt 2>&1 &
修改执行权限:
chmod +x ./start_maxwell.sh向MySQL中增删改查写入数据
mysql> create database mysqldb; Query OK, 1 row affected (0.01 sec) mysql> use mysqldb; Database changed mysql> create table info(id int,name varchar(255),age int); Query OK, 0 rows affected (0.04 sec) mysql> insert into info values (10,"xx",20); Query OK, 1 row affected (0.01 sec) mysql> insert into info values (12,"xx",20); Query OK, 1 row affected (0.04 sec) mysql> update info set age = 100 where id = 10; Query OK, 1 row affected (0.01 sec) Rows matched: 1 Changed: 1 Warnings: 0 mysql> delete from info where id = 10; Query OK, 1 row affected (0.00 sec)
对应Kafka中的消息如下:
[root@node3 bin]# ./kafka-console-consumer.sh --bootstrap-server node1:9092,node2:9092,node3:9092 --topic maxwell_topic {"database":"mysqldb","table":"info","type":"insert","ts":1641292268,"xid":2568,"commit":true,"data":{"id":12,"name":"xx","age":20}} {"database":"mysqldb","table":"info","type":"update","ts":1641292291,"xid":2631,"commit":true,"data":{"id":10,"name":"xx","age":100},"old":{"age":20}} {"database":"mysqldb","table":"info","type":"delete","ts":1641292302,"xid":2662,"commit":true,"data":{"id":10,"name":"xx","age":100}}测试Maxwell断点续传
停止Maxwell,向MySQL中插入新的数据,重启Maxwell观察是否从上次消费到的binlog位置继续消费。
#使用kill -9 xxx 命令在node3停止Maxwell #向MySQL继续插入数据 mysql> insert into info values (200,"bbb",20); mysql> update info set age = 30 where id = 200;
#重新在node3启动Maxwell,可以观察到Kafka中继续上次binlog位置写入数据 {"database":"mysqldb","table":"info","type":"insert","ts":1619000378,"xid":4565,"commit":true,"data" :{"id":200,"name":"bbb","age":20}}{"database":"mysqldb","table":"info","type":"update","ts":1619000391,"xid":4566,"commit":true,"data" :{"id":200,"name":"bbb","age":30},"old":{"age":20}}
欢迎分享,转载请注明来源:内存溢出
评论列表(0条)