debezium 依赖于kafka，kafka依赖于zookeeper。
zookeeper实现了kafka消息的一致性，debezium 把订阅的数据推送到kafka

dockerFile

FROM debezium/connect:1.6
ENV KAFKA_CONNECT_JDBC_DIR=$KAFKA_CONNECT_PLUGINS_DIR/kafka-connect-jdbc
ENV INSTANT_CLIENT_DIR=/instant_client/

USER root

USER kafka
# Deploy Oracle client and drivers

COPY instant_client/* $INSTANT_CLIENT_DIR
COPY instant_client/xstreams.jar /kafka/libs
COPY instant_client/ojdbc8.jar /kafka/libs

INSTANT_CLIENT_DIR 为oracleCilent，去官网下载即可，oracle11g的client是jdk8使用的，而debezium是jdk11， 这里我使用的是oracle12的client，测试可以使用

docker-compose 可根据github的推荐进行修改
https://github.com/debezium/debezium-examples/tree/master/tutorial

docker-compose.yaml

version: '2'
services:
  zookeeper:
    image: debezium/zookeeper:1.6
    ports:
     - 2181:2181
     - 2888:2888
     - 3888:3888
    volumes:
     - D://software\docker-volume\docker_debezium/volume:/volume
    privileged: true
    networks:
      proxy:
        ipv4_address: 192.168.192.3
  kafka:
    depends_on: 
      - zookeeper
    image: debezium/kafka:1.6
    ports:
     - 9092:9092
    links:
     - zookeeper
    environment:
     - ZOOKEEPER_CONNECT=zookeeper:2181
     - cluster.id=012344
    volumes:
     - D://software\docker-volume\docker_debezium/volume:/volume
     - D://software\docker-volume\docker_debezium/kafka/data:/kafka/data
     - D://software\docker-volume\docker_debezium/kafka/logs:/kafa/logs
    cap_add:
     - ALL # 开启全部权限
    privileged: true #设置容器的权限为root
    networks:
      proxy:
        ipv4_address: 192.168.192.4
  connect:
    depends_on: 
      - kafka
    image: debezium-connenct-oracle
    build:
      context: D:\software\docker-volume\docker_debezium
      dockerfile: dockerFile
      args:
        DEBEZIUM_VERSION: 1.6
    ports:
     - 8083:8083
     - 5005:5005
    links:
     - kafka
    environment:
     - BOOTSTRAP_SERVERS=kafka:9092
     - GROUP_ID=1
     - CONFIG_STORAGE_TOPIC=my_connect_configs
     - OFFSET_STORAGE_TOPIC=my_connect_offsets
     - STATUS_STORAGE_TOPIC=my_connect_statuses
     - LD_LIBRARY_PATH=/instant_client
     - KAFKA_DEBUG=true
     - DEBUG_SUSPEND_FLAG=n
     - JAVA_DEBUG_PORT=0.0.0.0:5005
    volumes:
     - D://software\docker-volume\docker_debezium/volume:/volume 
    privileged: true
    networks:
      proxy:
        ipv4_address: 192.168.192.5
  kafkaui:
    depends_on: 
      - kafka
    image: provectuslabs/kafka-ui:latest
    ports:
     - 8811:8080
    links:
     - kafka
    environment:
     - KAFKA_CLUSTERS_0_NAME=clusters
     - KAFKA_CLUSTERS_0_BOOTSTRAPSERVERS=kafka:9092
    networks:
      proxy:
        ipv4_address: 192.168.192.6
     

networks:
  proxy:
    ipam:
      config:
      - subnet: 192.168.192.0/20

注意要将kafka的data映射出去，因为kafka容器重新构建会丢失未消费的消息。
所有volumes的映射根据自己需求修改
kafka的cluster.id最好是随意写一个，这里把所有容器的网络ip都定义好，方便后面使用。
容器启动后即可访问kafkaUi

oracle执行下列语句用以支持订阅

ALTER TABLE 表名 ADD SUPPLEMENTAL LOG DATA (ALL) COLUMNS
ALTER DATABASE ADD SUPPLEMENTAL LOG DATA

此时的kafka地址即为上面设置的192.168.192.4
使用postman或者curl给debezium发送创建连接器

http://localhost:8083/connectors/

{
  "name": "oracle-connector",
  "config": {
    "connector.class": "io.debezium.connector.oracle.OracleConnector",
    "database.hostname": "xxx",
    "database.port": "1521",
    "database.user": "xxx",
    "database.password": "xxx",
    "database.server.name": "xxx",
    "database.history.kafka.bootstrap.servers": "192.168.192.4:9092",
    "database.history.kafka.topic": "oracle-history",
    "topic.prefix":"prefix",
    "database.dbname":"xxx",
    "table.include.list": "你订阅的表",
    "snapshot.mode": "schema_only",
    "database.history.skip.unparseable.ddl":"true"
  }
}

接口返回为你的配置即配置成功

使用
get http://localhost:8083/connectors/${connector}/status 获取连接器运行状态
如上则是 http://localhost:8083/connectors/oracle-connector/status

备注
“snapshot.mode”: "schema_only"表示不同步表所有信息，只同步当前增量数据。
“database.history.skip.unparseable.ddl”:“true” 在对ddl语句解析失败时忽略，程序很容易j解析ddl异常如

 Ignoring unparsable DDL statement 'alter table SYSTEM.LOGMNR_TABCOMPART$ modify partition P470 REBUILD UNUSABLE LOCAL INDEXES;': {}   [io.debezium.connector.oracle.OracleSchemaChangeEventEmitter]

io.debezium.text.ParsingException: DDL statement couldn't be parsed. Please open a Jira issue with the statement 'alter table SYSTEM.LOGMNR_TABCOMPART$ modify partition P470 REBUILD UNUSABLE LOCAL INDEXES;'

而异常后连接器会直接挂掉，配置忽略后异常不会挂而是忽略此错误

至此配置完成，进入kafkaui查看数据

有一个以你设置的前缀开头的和你表名结尾的队列即为你订阅的表，如上配置则为
prefix.你订阅的表
你对表 insert、update、delete的数据全部同步至此，消息包含before、after字段用以对比更改前后数据。

kafka 的配置修改 connect-distributed.properties

key.converter.schemas.enable=false
value.converter.schemas.enable=false

用以精简json

至此搭建已完毕。