华为FusionInsight MRS CDL使用指南

说明

CDL是一种简单、高效的数据实时集成服务，能够从各种OLTP数据库中抓取Data Change事件，然后推送至Kafka中，最后由Sink Connector消费Topic中的数据并导入到大数据生态软件应用中，从而实现数据的实时入湖。

CDL服务包含了两个重要的角色：CDLConnector和CDLService。CDLConnector是具体执行数据抓取任务的实例，CDLService是负责管理和创建任务的实例。

本此实践介绍以mysql作为数据源进行数据抓取

前提条件

• MRS集群已安装CDL服务。

• MySQL数据库需要开启mysql的bin log功能（默认情况下是开启的）。

  查看MySQL是否开启bin log：

  使用工具或者命令行连接MySQL数据库（本示例使用navicat工具连接），执行show variables like 'log_%'命令查看。

  例如在navicat工具选择"File > New Query"新建查询，输入如下SQL命令，单击"Run"在结果中"log_bin"显示为"ON"则表示开启成功。

  show variables like 'log_%'

  

工具准备

现在cdl只能使用rest api的方式进行命令提交，所以需要提前安装工具进行调试。本文使用VSCode工具。

完成之后安装rest client插件：

完成之后创建一个cdl.http的文件进行编辑：

创建CDL任务

CDL任务创建的流程图如下所示：

说明：需要先创建一个MySQL link, 在创建一个Kafka link, 然后再创建一个CDL同步任务并启动。

MySQL link部分rest请求代码

@hostname = 172.16.9.113
@port = 21495
@host = {{hostname}}:{{port}}
@bootstrap = "172.16.9.113:21007"
@bootstrap_normal = "172.16.9.113:21005"
@mysql_host = "172.16.2.118"
@mysql_port = "3306"
@mysql_database = "hudi"
@mysql_user = "root"
@mysql_password = "Huawei@123"

### get links
get https://{{host}}/api/v1/cdl/link

### mysql link validate

post https://{{host}}/api/v1/cdl/link?validate=true
content-type: application/json

{
"name": "MySQL_link", //link名，全局唯一，不能重复
"description":"MySQL connection", //link描述
"link-type":"mysql", //link的类型
"enabled":"true",
"link-config-values":  {
"inputs": [
        { "name": "host", "value": {{mysql_host}} }, //数据库安装节点的ip
        { "name": "port", "value": {{mysql_port}} },//数据库监听的端口
        { "name": "database.name", "value": {{mysql_database}} }, //连接的数据库名
        { "name": "user", "value": {{mysql_user}} }, //用户
        { "name": "password","value": {{mysql_password}} } ,//密码
        { "name":"schema", "value": {{mysql_database}}}//同数据库名
        ]
    }
}

### mysql link create

post https://{{host}}/api/v1/cdl/link
content-type: application/json

{
"name": "MySQL_link", //link名，全局唯一，不能重复
"description":"MySQL connection", //link描述
"link-type":"mysql", //link的类型
"enabled":"true",
"link-config-values":  {
"inputs": [
        { "name": "host", "value": {{mysql_host}} }, //数据库安装节点的ip
        { "name": "port", "value": {{mysql_port}} },//数据库监听的端口
        { "name": "database.name", "value": {{mysql_database}} }, //连接的数据库名
        { "name": "user", "value": {{mysql_user}} }, //用户
        { "name": "password","value": {{mysql_password}} } ,//密码
        { "name":"schema", "value": {{mysql_database}}}//同数据库名
        ]
    }
}

### mysql link update

put https://{{host}}/api/v1/cdl/link/MySQL_link
content-type: application/json

{
"name": "MySQL_link", //link名，全局唯一，不能重复
"description":"MySQL connection", //link描述
"link-type":"mysql", //link的类型
"enabled":"true",
"link-config-values":  {
"inputs": [
        { "name": "host", "value": {{mysql_host}} }, //数据库安装节点的ip
        { "name": "port", "value": {{mysql_port}} },//数据库监听的端口
        { "name": "database.name", "value": {{mysql_database}} }, //连接的数据库名
        { "name": "user", "value": {{mysql_user}} }, //用户
        { "name": "password","value": {{mysql_password}} } ,//密码
        { "name":"schema", "value": {{mysql_database}}}//同数据库名
        ]
    }
}

Kafka link部分rest请求代码

### get links
get https://{{host}}/api/v1/cdl/link

### kafka link validate

post https://{{host}}/api/v1/cdl/link?validate=true
content-type: application/json

{
"name": "kafka_link",
"description":"test kafka link",
"link-type":"kafka",
"enabled":"true",
"link-config-values":  {
"inputs": [
        { "name": "bootstrap.servers", "value": "172.16.9.113:21007" },
        { "name": "sasl.kerberos.service.name", "value": "kafka" },
        { "name": "security.protocol","value": "SASL_PLAINTEXT" }//安全模式为SASL_PLAINTEXT，普通模式为PLAINTEXT
        ]
    }
}

### kafka link create

post https://{{host}}/api/v1/cdl/link
content-type: application/json

{
"name": "kafka_link",
"description":"test kafka link",
"link-type":"kafka",
"enabled":"true",
"link-config-values":  {
"inputs": [
        { "name": "bootstrap.servers", "value": "172.16.9.113:21007" },
        { "name": "sasl.kerberos.service.name", "value": "kafka" },
        { "name": "security.protocol","value": "SASL_PLAINTEXT" }//安全模式为SASL_PLAINTEXT，普通模式为PLAINTEXT
        ]
    }
}

### kafka link update

put https://{{host}}/api/v1/cdl/link/kafka_link
content-type: application/json

{
"name": "kafka_link",
"description":"test kafka link",
"link-type":"kafka",
"enabled":"true",
"link-config-values":  {
"inputs": [
        { "name": "bootstrap.servers", "value": "172.16.9.113:21007" },
        { "name": "sasl.kerberos.service.name", "value": "kafka" },
        { "name": "security.protocol","value": "SASL_PLAINTEXT" }//安全模式为SASL_PLAINTEXT，普通模式为PLAINTEXT
        ]
    }
}

CDL任务命令部分rest请求代码

@hostname = 172.16.9.113
@port = 21495
@host = {{hostname}}:{{port}}
@bootstrap = "172.16.9.113:21007"
@bootstrap_normal = "172.16.9.113:21005"
@mysql_host = "172.16.2.118"
@mysql_port = "3306"
@mysql_database = "hudi"
@mysql_user = "root"
@mysql_password = "Huawei@123"

### create job
post https://{{host}}/api/v1/cdl/job
content-type: application/json

{
    "job_type": "CDL_JOB", //job类型,目前只支持CDL_JOB这一种
    "name": "mysql_to_kafka", //job名称
    "description":"mysql_to_kafka", //job描述
    "from-link-name": "MySQL_link",  //数据源Link
    "to-link-name": "kafka_link", //目标源Link
    "from-config-values": {
        "inputs": [
            {"name" : "connector.class", "value" : "com.huawei.cdc.connect.mysql.MysqlSourceConnector"},
            {"name" : "schema", "value" : "hudi"},
            {"name" : "db.name.alias", "value" : "hudi"},
            {"name" : "whitelist", "value" : "hudisource"},
            {"name" : "tables", "value" : "hudisource"},
            {"name" : "tasks.max", "value" : "10"},
            {"name" : "mode", "value" : "insert,update,delete"},
            {"name" : "parse.dml.data", "value" : "true"},
            {"name" : "schema.auto.creation", "value" : "false"},
            {"name" : "errors.tolerance", "value" : "all"},
            {"name" : "multiple.topic.partitions.enable", "value" : "false"},
            {"name" : "topic.table.mapping", "value" : "[
                    {\"topicName\":\"huditableout\", \"tableName\":\"hudisource\"}
                ]"
            },
  	        {"name" : "producer.override.security.protocol", "value" : "SASL_PLAINTEXT"},//安全模式为SASL_PLAINTEXT，普通模式为PLAINTEXT
	        {"name" : "consumer.override.security.protocol", "value" : "SASL_PLAINTEXT"}//安全模式为SASL_PLAINTEXT，普通模式为PLAINTEXT
        ]
    },
    "to-config-values": {"inputs": []},
    "job-config-values": {
        "inputs": [
            {"name" : "global.topic", "value" : "demo"}
        ]
    }
}

### get all job
get https://{{host}}/api/v1/cdl/job


### submit job
put https://{{host}}/api/v1/cdl/job/mysql_to_kafka/start


### get job status
get https://{{host}}/api/v1/cdl/submissions?jobName=mysql_to_kafka


### stop job
put https://{{host}}/api/v1/cdl/job/mysql_to_kafka/submissions/13/stop


### delete job
DELETE https://{{host}}/api/v1/cdl/job/mysql_to_kafka

场景验证

生产库MySQL原始数据如下：

提交CDL任务之后

增加操作: insert into hudi.hudisource values (11,“蒋语堂”,38,“女”,“图”,“播放器”,28732);

对应kafka消息体：

更改操作: UPDATE hudi.hudisource SET uname=‘Anne Marie333’ WHERE uid=11;

对应kafka消息体：

删除操作：delete from hudi.hudisource where uid=11;

对应kafka消息体：