Debezium的ExtractNewRecordState转换无法工作

Question

我正在构建一个数据同步器，它捕获来自MySQL源的数据更改，并将数据导出到配置单元。

我选择使用Kafka Connect来实现此功能。我使用Debezium作为源连接器，使用confluent hdfs作为宿连接器。

Debezium提供了一个single message transformation，让我可以从复杂的事件消息中提取after字段。我执行了与列出的文档相同的配置，但它不起作用。

{
    // omit ...
    "transform": "unwrap",
    "transform.unwrap.type": "io.debezium.transforms.ExtractNewRecordState"
}

我已经尝试在源连接器端和接收器连接器端配置转换，它仍然不能工作。事实上，当我在源连接器端配置它，然后检查相应主题中的消息时，我发现消息仍然包含所有字段，包括before、source等。

ythh@openstack2:~/confluent-5.5.0$ bin/kafka-avro-console-consumer --from-beginning --bootstrap-server localhost:9092 --topic dbserver1.test_data_1.student3
{"before":null,"after":{"dbserver1.test_data_1.student3.Value":{"id":1,"name":"ggg"}},"source":{"version":"1.1.1.Final","connector":"mysql","name":"dbserver1","ts_ms":1589005572000,"snapshot":{"string":"false"},"db":"test_data_1","table":{"string":"student3"},"server_id":1,"gtid":null,"file":"mysql-bin.000011","pos":9474,"row":0,"thread":{"long":6013},"query":null},"op":"c","ts_ms":{"long":1589005572172},"transaction":null}
{"before":null,"after":{"dbserver1.test_data_1.student3.Value":{"id":2,"name":"no way"}},"source":{"version":"1.1.1.Final","connector":"mysql","name":"dbserver1","ts_ms":1589005893000,"snapshot":{"string":"false"},"db":"test_data_1","table":{"string":"student3"},"server_id":1,"gtid":null,"file":"mysql-bin.000011","pos":11218,"row":0,"thread":{"long":6030},"query":null},"op":"c","ts_ms":{"long":1589005893773},"transaction":null}
{"before":null,"after":{"dbserver1.test_data_1.student3.Value":{"id":3,"name":"not work"}},"source":{"version":"1.1.1.Final","connector":"mysql","name":"dbserver1","ts_ms":1589005900000,"snapshot":{"string":"false"},"db":"test_data_1","table":{"string":"student3"},"server_id":1,"gtid":null,"file":"mysql-bin.000011","pos":11501,"row":0,"thread":{"long":6030},"query":null},"op":"c","ts_ms":{"long":1589005900724},"transaction":null}

我还检查了kafka连接日志，下面是一些输出：

ythh@openstack2:~/kafka_2.12-2.5.0/logs$ cat connect.log | grep transform
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
[2020-05-09 14:29:30,470] INFO    transform.unwrap.type = io.debezium.transforms.ExtractNewRecordState (io.debezium.connector.common.BaseSourceTask:97)
[2020-05-09 14:29:30,470] INFO    transform = unwrap (io.debezium.connector.common.BaseSourceTask:97)
[2020-05-09 14:29:30,471] INFO    transform.unwrap.drop.tombstones = false (io.debezium.connector.common.BaseSourceTask:97)
[2020-05-09 14:29:30,471] INFO    transform.unwrap.delete.handling.mode = rewrite (io.debezium.connector.common.BaseSourceTask:97)
        transforms = []
        transforms = []
[2020-05-09 14:29:32,419] INFO    transform.unwrap.type = io.debezium.transforms.ExtractNewRecordState (io.debezium.connector.common.BaseSourceTask:97)
[2020-05-09 14:29:32,419] INFO    transform = unwrap (io.debezium.connector.common.BaseSourceTask:97)
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []
        transforms = []

Answer 1

看起来你打错了(transform而不是transforms)。尝试此配置：

{
    // omit ...
    "transforms": "unwrap",
    "transforms.unwrap.type": "io.debezium.transforms.ExtractNewRecordState"
}