apache-kafka - 从主题创建表并得到一些关于 LongDeserializer 接收的数据大小不是 8 的序列化异常

Question

环境(Docker):

# 5.5.1
image: confluentinc/cp-zookeeper:latest

# 2.13-2.6.0
image: wurstmeister/kafka:latest

# 5.5.1
image: confluentinc/cp-schema-registry:latest

# 5.5.1
image: confluentinc/cp-kafka-connect:latest

# 0.11.0
image: confluentinc/ksqldb-server:latest

Kafka 主题内容来自 kafka connect(使用 debezium)。
当我使用查询( select * from user emit changes )时，
显示了大部分内容，但丢失了一些内容。
我尝试查看来自 ksqldb-server 的日志，并发现了一些错误消息:

ksqldb-server      | [2020-08-29 12:44:23,008] ERROR {"type":0,"deserializationError":{"errorMessage":"Error deserializing DELIMITED message from topic: pa.new_pa.user","recordB64":null,"cause":["Size of data received by LongDeserializer is not 8"],"topic":"pa.new_pa.user"},"recordProcessingError":null,"productionError":null} (processing.CTAS_USER2_0.KsqlTopic.Source.deserializer:44)
ksqldb-server      | [2020-08-29 12:44:23,008] WARN Exception caught during Deserialization, taskId: 0_0, topic: pa.new_pa.user, partition: 0, offset: 23095 (org.apache.kafka.streams.processor.internals.StreamThread:36)
ksqldb-server      | org.apache.kafka.common.errors.SerializationException: Error deserializing DELIMITED message from topic: pa.new_pa.user
ksqldb-server      | Caused by: org.apache.kafka.common.errors.SerializationException: Size of data received by LongDeserializer is not 8
ksqldb-server      | [2020-08-29 12:44:23,009] WARN stream-thread [_confluent-ksql-default_query_CTAS_USER2_0-6637e2a8-c417-49fa-bb65-d0d1a5205af1-StreamThread-1] task [0_0] Skipping record due to deserialization error. topic=[pa.new_pa.user] partition=[0] offset=[23095] (org.apache.kafka.streams.processor.internals.RecordDeserializer:88)
ksqldb-server      | org.apache.kafka.common.errors.SerializationException: Error deserializing DELIMITED message from topic: pa.new_pa.user
ksqldb-server      | Caused by: org.apache.kafka.common.errors.SerializationException: Size of data received by LongDeserializer is not 8

我尝试使用偏移量为“23095”的消息，看起来不错。

[2020-08-29 13:24:12,021] INFO [Consumer clientId=consumer-console-consumer-37294-1, groupId=console-consumer-37294] Subscribed to partition(s): pa.new_pa.user-0 (org.apache.kafka.clients.consumer.KafkaConsumer)
[2020-08-29 13:24:12,026] INFO [Consumer clientId=consumer-console-consumer-37294-1, groupId=console-consumer-37294] Seeking to offset 23095 for partition pa.new_pa.user-0 (org.apache.kafka.clients.consumer.KafkaConsumer)
[2020-08-29 13:24:12,570] INFO [Consumer clientId=consumer-console-consumer-37294-1, groupId=console-consumer-37294] Cluster ID: rdsgvpoESzer6IAxQDlLUA (org.apache.kafka.clients.Metadata)
{"id":8191,"parent_id":{"long":8184},"upper_id":0,"username":"app0623c","domain":43,"role":1,"modified_at":1598733553000,"blacklist_modified_at":{"long":1598733768000},"tied_at":{"long":1598733771000},"name":"test","enable":1,"is_default":0,"bankrupt":0,"locked":0,"tied":0,"checked":0,"failed":0,"last_login":{"long":1598733526000},"last_online":{"long":1598733532000},"last_ip":{"bytes":"ÿÿ\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000"},"last_country":{"string":"MY"},"last_city_id":0}

这是我的源连接器配置，其他:

CREATE SOURCE CONNECTOR `pa_source_unwrap` WITH(
    "connector.class" = 'io.debezium.connector.mysql.MySqlConnector',
    "tasks.max" = '1',
    "database.hostname" = 'docker.for.mac.host.internal',
    "database.port" = '3306',
    "database.user" = 'root',
    "database.password" = 'xxxxxxx',
    "database.service.id" = '10001',
    "database.server.name" = 'pa',
    "database.whitelist" = 'new_pa',
    "table.whitelist" = 'new_pa.user, new_pa.user_created, new_pa.cash',
    "database.history.kafka.bootstrap.servers" = 'kafka:9092',
    "database.history.kafka.topic" = 'schema-changes.pa',
    "transforms" = 'unwrap',
    "transforms.unwrap.type" = 'io.debezium.transforms.ExtractNewRecordState',
    "transforms.unwrap.delete.handling.mode" = 'drop',
    "transforms.unwrap.drop.tombstones" = 'true',
    "key.converter" = 'io.confluent.connect.avro.AvroConverter',
    "value.converter" = 'io.confluent.connect.avro.AvroConverter',
    "key.converter.schema.registry.url" = 'http://schema-registry:8081',
    "value.converter.schema.registry.url" = 'http://schema-registry:8081',
    "key.converter.schemas.enable" = 'true',
    "value.converter.schemas.enable" = 'true'
);

CREATE TABLE user (`id` BIGINT PRIMARY KEY) WITH (
    KAFKA_TOPIC = 'pa.new_pa.user',
    VALUE_FORMAT = 'AVRO'
);

主题架构(自动生成):

Key:
{
  "connect.name": "pa.new_pa.user.Key",
  "fields": [
    {
      "name": "id",
      "type": "long"
    }
  ],
  "name": "Key",
  "namespace": "pa.new_pa.user",
  "type": "record"
}

Value:
{
  "connect.name": "pa.new_pa.user.Value",
  "fields": [
    {
      "name": "id",
      "type": "long"
    },
    {
      "default": null,
      "name": "parent_id",
      "type": [
        "null",
        "long"
      ]
    },
    {
      "default": 0,
      "name": "upper_id",
      "type": {
        "connect.default": 0,
        "type": "long"
      }
    },
    {
      "name": "username",
      "type": "string"
    },
    {
      "name": "domain",
      "type": "int"
    },
    {
      "name": "role",
      "type": {
        "connect.type": "int16",
        "type": "int"
      }
    },
    {
      "name": "modified_at",
      "type": {
        "connect.name": "io.debezium.time.Timestamp",
        "connect.version": 1,
        "type": "long"
      }
    },
    {
      "default": null,
      "name": "blacklist_modified_at",
      "type": [
        "null",
        {
          "connect.name": "io.debezium.time.Timestamp",
          "connect.version": 1,
          "type": "long"
        }
      ]
    },
    {
      "default": null,
      "name": "tied_at",
      "type": [
        "null",
        {
          "connect.name": "io.debezium.time.Timestamp",
          "connect.version": 1,
          "type": "long"
        }
      ]
    },
    {
      "name": "name",
      "type": "string"
    },
    {
      "name": "enable",
      "type": {
        "connect.type": "int16",
        "type": "int"
      }
    },
    {
      "name": "is_default",
      "type": {
        "connect.type": "int16",
        "type": "int"
      }
    },
    {
      "name": "bankrupt",
      "type": {
        "connect.type": "int16",
        "type": "int"
      }
    },
    {
      "name": "locked",
      "type": {
        "connect.type": "int16",
        "type": "int"
      }
    },
    {
      "name": "tied",
      "type": {
        "connect.type": "int16",
        "type": "int"
      }
    },
    {
      "name": "checked",
      "type": {
        "connect.type": "int16",
        "type": "int"
      }
    },
    {
      "name": "failed",
      "type": {
        "connect.type": "int16",
        "type": "int"
      }
    },
    {
      "default": null,
      "name": "last_login",
      "type": [
        "null",
        {
          "connect.name": "io.debezium.time.Timestamp",
          "connect.version": 1,
          "type": "long"
        }
      ]
    },
    {
      "default": null,
      "name": "last_online",
      "type": [
        "null",
        {
          "connect.name": "io.debezium.time.Timestamp",
          "connect.version": 1,
          "type": "long"
        }
      ]
    },
    {
      "default": null,
      "name": "last_ip",
      "type": [
        "null",
        "bytes"
      ]
    },
    {
      "default": null,
      "name": "last_country",
      "type": [
        "null",
        "string"
      ]
    },
    {
      "name": "last_city_id",
      "type": "long"
    }
  ],
  "name": "Value",
  "namespace": "pa.new_pa.user",
  "type": "record"
}

Answer 1

这里的问题是您的 key 在 Avro 中，而 ksqlDB 目前仅支持 KAFKA 格式的 key (从 0.12 版开始)。
Avro key 正在积极处理中:#4461添加了对 Avro 原语和 #4997 的支持扩展它以支持 Avro 记录中的单个键列，(如您所见)。
您正在使用配置设置 Avro key 格式:

"key.converter" = 'io.confluent.connect.avro.AvroConverter',

你是 SQL:

CREATE TABLE user (`id` BIGINT PRIMARY KEY) WITH (
    KAFKA_TOPIC = 'pa.new_pa.user',
    VALUE_FORMAT = 'AVRO'
);

正在设置 VALUE_FORMAT至 AVRO ，但目前 key 格式为 KAFKA .因此，您可以使用:

"key.converter" = 'org.apache.kafka.connect.converters.IntegerConverter',

...将 key 转换为正确的格式。有关用于 KAFKA 格式的正确转换器的更多信息，请访问 ksqlDB microsite .