通过Debezium连接器获取sqlserver实时数据并传入kafka

通过Debezium连接器获取sqlserver实时数据

1、下载连接器

debezium-connector-sqlserver-1.2.5.Final-plugin.tar.gz 下载地址

2、配置kafka Distributed模式启动

在启动kafkaconnect的distributed模式之前，首先需要创建三个主题，三个主题分别对应connect-distributed.propertites文件中config.storage.topic(default connect-configs)、offset.storage.topic (default connect-offsets) 、status.storage.topic (default connect-status)的配置。

1、config.storage.topic：用以保存connector和task的配置信息，需要注意的是这个主题的分区数只能是1，而且是有多副本的。
2、offset.storage.topic:用以保存offset信息。
3、status.storage.topic:用以保存connetor的状态信息。

2.1 创建topic

./kafka-topics.sh --zookeeper 10.170.130.xxx:2181 --create --topic connect-offsets --replication-factor 2 --partitions 3
./kafka-topics.sh --zookeeper 10.170.130.xxx:2181 --create --topic connect-configs --replication-factor 2 --partitions 1
./kafka-topics.sh --zookeeper 10.170.130.xxx:2181 --create --topic connect-status --replication-factor 2 --partitions 3

replication-factor 副本数
partitions 也根据集群规模设定

可能用到的命令：

kafka-topics --list --zookeeper 10.2.13.xxx:2181
kafka-topics --create --zookeeper 10.2.13.xxx:2181  --replication-factor 1 --partitions 3 --topic topic_calllog_phone6
kafka-topics --delete --topic topic_ArrayVPN_Log --zookeeper 10.160.14.xxx:2181
kafka-console-producer --broker-list 10.2.13.xxx:9092 --topic topic_calllog_phone
kafka-console-consumer --bootstrap-server 10.2.13.xx:9092 --topic topic_calllog_phone6 --from-beginning

3、将需要的connector jar包解压到connectors文件夹下（每个节点）

cd /usr/local/kafka
mkdir connectors

#上传debezium-connector-sqlserver-1.2.5.Final-plugin.tar.gz 到服务器上
tar -zxvf debezium-connector-sqlserver-1.2.5.Final-plugin.tar.gz  -C     /usr/local/kafka/connectors/

4、编写connect-distributed.properties配置文件

cd /usr/local/kafka/config/
# kafka集群地址
bootstrap.servers=10.170.130.xxx:9092,10.170.130.xxx:9092,10.170.130.xxx:9092

# Connector集群的名称，同一集群内的Connector需要保持此group.id一致
group.id=connect-cluster

# 存储到kafka的数据格式
key.converter=org.apache.kafka.connect.json.JsonConverter
value.converter=org.apache.kafka.connect.json.JsonConverter
key.converter.schemas.enable=false
value.converter.schemas.enable=false

# 内部转换器的格式，针对offsets、config和status，一般不需要修改（一定要手动添加）
internal.key.converter=org.apache.kafka.connect.json.JsonConverter
internal.value.converter=org.apache.kafka.connect.json.JsonConverter
internal.key.converter.schemas.enable=false
internal.value.converter.schemas.enable=false

# 用于保存offsets的topic，应该有多个partitions，并且拥有副本(replication)
# Kafka Connect会自动创建这个topic，但是你可以根据需要自行创建
offset.storage.topic=connect-offsets
offset.storage.replication.factor=3
offset.storage.partitions=3


# 保存connector和task的配置，应该只有1个partition，并且有多个副本
config.storage.topic=connect-configs
config.storage.replication.factor=3

# 用于保存状态，可以拥有多个partition和replication
status.storage.topic=connect-status
status.storage.replication.factor=3
status.storage.partitions=3

# Flush much faster than normal, which is useful for testing/debugging
offset.flush.interval.ms=10000

# RESET主机名，默认为本机
#rest.host.name=
# REST端口号
rest.port=18083

# The Hostname & Port that will be given out to other workers to connect to i.e. URLs that are routable from other servers.
#rest.advertised.host.name=
#rest.advertised.port=

# 保存connectors的路径
#plugin.path=/usr/local/share/java,/usr/local/share/kafka/plugins,/opt/connectors,
plugin.path=/opt/module/kafka_2.11-2.4.0/connectors

5、分发配置到其余节点

6、启动kafka connect集群（每台节点执行命令）

#开启守护进程并启动
./connect-distributed.sh  -daemon /usr/local/kafka/config/connect-distributed.properties

6.1 页面访问

http://10.170.130.xxx:18083/
返回：
{"version":"2.2.2","commit":"1d348535a0a747d1","kafka_cluster_id":"9gckc_g_Q-eOFjTHSLMacA"}

http://hserver1:18083/connectors/  
返回：【】

curl -s hserver1:18083/connector-plugins | jq
返回：plugins下放的jar包

---

7、配置sqlserver CDC数据变更模式

7.1启动serveragent代理服务（已启动，启动步骤在sqlserver笔记本中）

7.2 开启数据库以及表的CDC模式

#创建数据库
Create database oneJ

#开启数据库的CDC

USE oneJ 
GO 
EXEC sys.sp_cdc_enable_db 
GO

# 开启表的CDC

USE oneJ 
GO 
EXEC sys.sp_cdc_enable_table 
@source_schema = N'dbo', 
@source_name = N'bs_year_task_gather_unit', 
@role_name = NULL, 
@supports_net_changes = 1 
GO

8、 提交Connector用户配置

提交之前要手动创建以下参数中的kafka的topic

kafka-topics --create --zookeeper 10.2.13.xxx:2181  --replication-factor 2 --partitions 3 --topic dbhistory.fullfillment

暂时用的psotman提交成功，以下方式还没测试

# curl -s -X POST -H "Content-Type: application/json" --data '{

{
  "name": "inventory-connector",  
  "config": {
    "connector.class": "io.debezium.connector.sqlserver.SqlServerConnector", 
    "database.hostname": "10.170.130.xxx", 
    "database.port": "1433", 
    "database.user": "sa", 
    "database.password": "123456", 
    "database.dbname": "testme", 
    "database.server.name": "fullfillment", 
    "table.whitelist": "dbo.bs_news_docs", 
    "database.history.kafka.bootstrap.servers": "10.170.130.xxx:9092,10.170.130.xxx:9092,10.170.130.xxx:9092", 
    "database.history.kafka.topic": "dbhistory.fullfillment" 
  }
}

}' http://10.170.130.xxx:18083/connectors | jq

上述参数说明
1、name:自定义连接器名称
2、database.server.name:自定义名字（在kafka生成的topic前缀相关）
3、table.whitelist：监听的表列表
4、database.history.kafka.topic：需要存储信息的topic

8.1 其余的一些操作

#查看connector的信息
 curl -s hserver1:18083/connectors/connector-mssql-online | jq
 
 #查看connector下运行的task信息
 curl -s hserver1:18083/connectors/connector-mssql-online/tasks | jq
 
 #查看connector当前状态
 
#curl -s hserver1:18083/connectors/connector-mssql-online/status | jq

#删除 Connector
curl -s -X DELETE hserver1:18083/connectors/connector-mssql-online


#暂停/重启 Connector
 curl -s -X PUT hserver1:18083/connectors/connector-mssql-online/pause
 curl -s -X PUT hserver1:18083/connectors/connector-mssql-online/resume

出现监听不了的状况
将库的监听释放
重新监听表

cd /
find -name "sqlagent.out"

– 对表禁用变更数据捕获
USE MyDatabase;
GO

EXEC sys.sp_cdc_disable_table
@source_schema = N ‘dbo’ ,
@source_name = N ‘CDC_Test’ ,
@capture_instance = N ‘dbo_CDC_Test’
GO

– 对数据库禁用变更数据捕获
USE MyDatabase;
GO
EXECUTE sys.sp_cdc_disable_db;
GO

其余参考地址