备注:我本地目录规划,统一放到文件夹:
- D:\java\Projects\sync-data\datax
- D:\java\Projects\sync-data\datax-web
一、安装DataX
(1) 安装python3
(配置环境变量),但从github上下载的版本只支持python2.x版本,而现在python3.x正流行,这给我们带来不少麻烦。
datax 支持Python3需要修改替换datax/bin下面的三个python文件,替换文件地址:https://github.com/WeiYe-Jing/datax-web/tree/master/doc/datax-web/datax-python3
(2) 安装jdk8
(配置环境变量)
(3) 安装 maven 3.x
(配置环境变量)
(4) datax.tar.gz
包下载并解压:http://datax-opensource.oss-cn-hangzhou.aliyuncs.com/datax.tar.gz
(5) datax
源码下载:https://github.com/alibaba/DataX。(下载源码主要是因为某些源码需要自己修改,然后重新打包覆盖导datax,如下述的mysql版本不一致例子)
- ps1:自动下载依赖,如果
hitsdb-client
报错,则将hitsdb-client
依赖版本改为0.3.7即可 - ps2:因为我的项目用的是mysql8,所以需要将
mysqlReader
、mysqlWriter
服务的mysql
的依赖改为8.0.17
,默认是5.x
- ps3:
datax-all
服务执行clean-install
- ps4:将
mysqlReader
、mysqlWriter
生成的target
–>datax
–>plugins
复制到datax
的bin
同级的目录下,覆盖即可 - ps5:将“
datax\plugin\reader\mysqlreader\libs
”、“datax\plugin\writer\mysqlwriter\libs
”目录下的mysql5
的包删除
二、使用DataX
(1) 打开cmd界面
(2) 进入datax
的bin
目录下
(3) 输入CHCP 65001
防止中文乱码
(4) 查看模板命令行:python datax.py -r streamreader -w streamwriter
(5) 执行脚本:python datax.py E:\datax\datax\datax\job\job.json
ps1:报错:提供的配置文件[D:\java\Projects\sync-data\datax\plugin\reader._cassandrareader\plugin.json]不存在. 请检查您的配置文件.
—>解决:删除datax\plugin\writer、datax\plugin\reader底下"._"开头文件
(6) 编写脚本,将脚本存放到datax\datax\job
目录底下,然后执行后python datax.py datax\job\你的脚本.json
运行
ps1:脚本格式信息可以去https://github.com/alibaba/DataX
/查看
ps2:脚本基础参数说明(mysql、oracle)
{ "job": { "setting": { "speed": { "channel": 并发数 } }, "content": [ { "reader": { "name": "固定的名字", "parameter": { "username": "账号", "password": "密码", "column": [ "字段1", "字段2", "字段3" ps:[*]表示所有列,但是不建议使用 ], "splitPk": "分片字段,丢给task,所以必须是整形。一般用主键即可(非必须)", "connection": [ { "table": [ "表"], "jdbcUrl": [ "数据库连接地址"], "querySql":["这里可以写查询数据的sql语句,但是如果这里写了,则不允许再配置table,否则会报错(非必须,table和querySql只能存在一个)"] } ], "where":"过滤条件" } }, "writer": { "name": "固定的名字", "parameter": { "writeMode": "写入策略", "username": "账户", "password": "密码", "column": [ "字段1", "字段2", "字段3" ], "session": [ "DataX在获取Mysql连接时,执行session指定的SQL语句,修改当前connection session属性(非必须)" ], "preSql": [ "写入数据到目的表前,会先执行这里的标准语句(非必须)" ], "connection": [ { "jdbcUrl": "数据库连接地址", "table": [ "表" ] } ] } } } ] } }
ps3:将mysql的datax数据过滤后迁移到oracle实例
{ "job": { "setting": { "speed": { "channel": 1 } }, "content": [ { "reader": { "name": "mysqlreader", "parameter": { "username": "root", "password": "tiger", "column": [ "id", "name", "time" ], "splitPk": "id", "connection": [ { "table": [ "datax" ], "jdbcUrl": [ "jdbc:mysql://127.0.0.1:3306/datax?useUnicode=true&characterEncoding=utf-8&allowMultiQueries=true&useSSL=false&serverTimezone=GMT%2b8" ] } ], "where":"id < 10" } }, "writer": { "name": "oraclewriter", "parameter": { "username": "HYDROPOWER_JIANGXI", "password": "ffcsict123", "column": [ "id", "name", "time" ], "preSql": [ "delete from datax" ], "connection": [ { "jdbcUrl": "jdbc:oracle:thin:@192.168.35.9:1521:orcl", "table": [ "datax" ] } ] } } } ] } }
(7) 达梦数据库:因为数据库介绍没有dm数据库,所以只能使用“通用RDBMS(支持所有关系型数据库)”这种模式进行传输。
{ "job": { "setting": { "speed": { "channel": 1 } }, "content": [ { "reader": { "name": "mysqlreader", "parameter": { "username": "root", "password": "tiger", "column": [ "id", "name", "time" ], "splitPk": "id", "connection": [ { "table": [ "datax" ], "jdbcUrl": [ "jdbc:mysql://127.0.0.1:3306/datax?useUnicode=true&characterEncoding=utf-8&allowMultiQueries=true&useSSL=false&serverTimezone=GMT%2b8" ] } ], "where":"id < 10" } }, "writer": { "name": "rdbmswriter", "parameter": { "username": "AES", "password": "ffcsict123", "column": [ "id", "name", "time" ], "preSql": [ "delete from datax" ], "connection": [ { "jdbcUrl": "jdbc:dm://127.0.0.1:5236/AES?zeroDateTimeBehavior=converToNull&useUnicode=true&characterEncoding=utf-8", "table": [ "datax" ] } ] } } } ] } }
(8) 文件流:只能读取规则的二维表结构文件,并且只能读取本地文件及远程服务器的文件,只能调用api接口获取流
{ "setting": {}, "job": { "setting": { "speed": { "channel": 2 } }, "content": [ { "reader": { "name": "ftpreader", "parameter": { "protocol": "sftp", "host": "192.168.248.11", "port": 22, "username": "root", "password": "ffcsict123", "path": [ "/root/datax/datax.txt","/root/datax/datax.csv" ], "column": [ { "index": 0, "type": "string" }, { "index": 1, "type": "string" }, { "index": 2, "type": "string", } ], "encoding": "GBK", "fieldDelimiter": ",", "skipHeader": "true" } }, "writer": { "name": "mysqlwriter", "parameter": { "writeMode": "insert", "username": "root", "password": "tiger", "column": [ "id", "name", "time" ], "session": [ "set session sql_mode='ANSI'" ], "preSql": [ "delete from datax1" ], "connection": [ { "jdbcUrl": "jdbc:mysql://127.0.0.1:3306/datax?useUnicode=true&characterEncoding=utf-8&allowMultiQueries=true&useSSL=false&serverTimezone=GMT%2b8", "table": [ "datax1" ] } ] } } } ] } }
三、安装datax-web
3.1 下载 data-web源码:https://github.com/WeiYe-Jing/datax-web
3.2 创建数据库:脚本文件路径–>datax-web\bin\db
3.3 修改datax-admin下配置文件resources/application.yml
修改数据源配置,目前仅支持mysql
#数据源 datasource: username: root password: root url: jdbc:mysql://localhost:3306/datax_web?serverTimezone=Asia/Shanghai&useLegacyDatetimeCode=false&useSSL=false&nullNamePatternMatchesAll=true&useUnicode=true&characterEncoding=UTF-8 driver-class-name: com.mysql.jdbc.Driver
修改日志路径path
# 配置mybatis-plus打印slq日志 logging: level: com.wugui.datax.admin.mapper: error path: ./data/applogs/admin
修改邮件发送配置(不需要可以不修改)
# datax-web email mail: host: smtp.qq.com port: 25 username: xxx@qq.com password: xxx properties: mail: smtp: auth: true starttls: enable: true required: true socketFactory: class: javax.net.ssl.SSLSocketFactory
3.4 修改datax-executor下配置文件resources/application.yml
修改日志路径path
# log config logging: config: classpath:logback.xml path: ./data/applogs/executor/jobhandler
修改datax.job配置
datax: job: admin: ### datax-web admin address addresses: http://127.0.0.1:8080 executor: appname: datax-executor ip: port: 9999 ### job log path logpath: ./data/applogs/executor/jobhandler ### job log retention days logretentiondays: 30 executor: jsonpath: D:\java\Projects\sync-data\json pypath: D:\java\Projects\sync-data\datax\bin\datax.py
- admin.addresses datax_admin部署地址,如调度中心集群部署存在多个地址则用逗号分隔,执行器将会使用该地址进行"执行器心跳注册"和"任务结果回调";
- executor.appname 执行器AppName,每个执行器机器集群的唯一标示,执行器心跳注册分组依据;
- executor.ip 默认为空表示自动获取IP,多网卡时可手动设置指定IP,该IP不会绑定Host仅作为通讯实用;地址信息用于 "执行器注册" 和 "调度中心请求并触发任务";
- executor.port 执行器Server端口号,默认端口为9999,单机部署多个执行器时,注意要配置不同执行器端口;
- executor.logpath 执行器运行日志文件存储磁盘路径,需要对该路径拥有读写权限;
- executor.logretentiondays 执行器日志文件保存天数,过期日志自动清理, 限制值大于等于3时生效; 否则, 如-1, 关闭自动清理功能;
- executor.jsonpath datax json临时文件保存路径
- pypath DataX启动脚本地址,例如:xxx/datax/bin/datax.py
如果系统配置DataX环境变量(DATAX_HOME)logpath、jsonpath、pypath可不配,log文件和临时json存放在环境变量路径下。
四、启动项目
(1) 本地idea开发环境
- 运行datax-admin下 DataXAdminApplication
- 运行datax-executor下 DataXExecutorApplication
admin启动成功后日志会输出三个地址,两个接口文档地址,一个前端页面地址
(2) Linux环境部署
- 1.本地安装好maven环境,安装此处细节忽略
- 2.执行mvn package -Dmaven.test.skip=true
- 3.打包成功后分别将datax-admin、datax-executor模块target下datax-admin-2.1.1.jar、datax-executor-2.1.1.jar放到指定目录
- 4.分别启动datax-admin-1.0.0.jar、datax-executor-1.0.0.jar
- 5.启动命令demo:
nohup java -Xmx1024M -Xms1024M -Xmn448M -XX:MaxMetaspaceSize=192M -XX:MetaspaceSize=192M -jar datax-admin-2.1.1.jar& nohup java -Xmx1024M -Xms1024M -Xmn448M -XX:MaxMetaspaceSize=192M -XX:MetaspaceSize=192M -jar datax-executor-2.1.1.jar&
五、启动成功
启动成功后打开页面(默认管理员用户名:admin 密码:123456),http://localhost:8080/index.html#/dashboard
六、集群部署
调度中心、执行器支持集群部署,提升调度系统容灾和可用性。
- 1.调度中心集群:
DB配置保持一致;
集群机器时钟保持一致(单机集群忽视); - 2.执行器集群:
执行器回调地址(admin.addresses)需要保持一致;执行器根据该配置进行执行器自动注册等操作。
同一个执行器集群内AppName(executor.appname)需要保持一致;调度中心根据该配置动态发现不同集群的在线执行器列表。
参考: