Flink-Exactly Once(如何保证数据的唯一性和不重复!)

Flink-Exactly Once(如何保证数据的唯一性和不重复!) kafka中如何保证数据不丢失

1. kafka中如何保证数据不丢失

2. 向kafka中sink数据

package com.wt.flink.sink
import org.apache.flink.api.common.serialization.SimpleStringSchema
import org.apache.flink.connector.kafka.sink.{KafkaRecordSerializationSchema, KafkaSink}
import org.apache.flink.streaming.api.scala._

object Demo5KafkaSink {

  def main(args: Array[String]): Unit = {
    val env: StreamExecutionEnvironment = StreamExecutionEnvironment.getExecutionEnvironment

    val studentDS: DataStream[String] = env.readTextFile("data/students.json")


    /**
     * 将数据保存到kafka中 --- kafka sink
     *
     * DeliveryGuarantee.EXACTLY_ONCE： 唯一一次
     * DeliveryGuarantee.AT_LEAST_ONCE： 至少一次，默认
     */


    val sink: KafkaSink[String] = KafkaSink
      .builder[String]()
      .setBootstrapServers("master:9092,node1:9092,node2:9092") //broker地址
      .setRecordSerializer(
        KafkaRecordSerializationSchema
          .builder[String]()
          .setTopic("students_json") //topic
          .setValueSerializationSchema(new SimpleStringSchema())
          .build())
      //.setDeliverGuarantee(DeliveryGuarantee.AT_LEAST_ONCE) //唯一一次
      .build()

    //使用kafka sink
    studentDS.sinkTo(sink)
    
    env.execute()
    //kafka-console-consumer.sh --bootstrap-server  master:9092,node2:9092,node2:9092 --from-beginning --topic students_json
  }
}

我们在从命令行读取学生的json数据

3 . Flink从kafka中读取数据

package com.wt.flink.core

import org.apache.flink.api.common.eventtime.WatermarkStrategy
import org.apache.flink.api.common.serialization.SimpleStringSchema
import org.apache.flink.connector.kafka.source.KafkaSource
import org.apache.flink.connector.kafka.source.enumerator.initializer.OffsetsInitializer
import org.apache.flink.runtime.state.hashmap.HashMapStateBackend
import org.apache.flink.streaming.api.CheckpointingMode
import org.apache.flink.streaming.api.environment.CheckpointConfig.ExternalizedCheckpointCleanup
import org.apache.flink.streaming.api.scala._

object Demo16ExactlyOnce {
  def main(args: Array[String]): Unit = {
    /**
     * 使用flink从kafka中读取数据，怎么保证数据处理的唯一一次
     *
     */
    val env: StreamExecutionEnvironment = StreamExecutionEnvironment.getExecutionEnvironment

    /**
     * 开启checkpoint
     *
     */
    // 每 1000ms 开始一次 checkpoint
    env.enableCheckpointing(20000)
    // 高级选项：
    // 设置模式为精确一次 (这是默认值)
    env.getCheckpointConfig.setCheckpointingMode(CheckpointingMode.EXACTLY_ONCE)

    // 确认 checkpoints 之间的时间会进行 500 ms
    env.getCheckpointConfig.setMinPauseBetweenCheckpoints(500)

    // Checkpoint 必须在一分钟内完成，否则就会被抛弃
    env.getCheckpointConfig.setCheckpointTimeout(60000)

    // 允许两个连续的 checkpoint 错误
    env.getCheckpointConfig.setTolerableCheckpointFailureNumber(2)

    // 同一时间只允许一个 checkpoint 进行
    env.getCheckpointConfig.setMaxConcurrentCheckpoints(1)

    // 使用 externalized checkpoints，这样 checkpoint 在作业取消后仍就会被保留
    //RETAIN_ON_CANCELLATION: 当任务取消时保留checkpoint
    env.getCheckpointConfig.setExternalizedCheckpointCleanup(
      ExternalizedCheckpointCleanup.RETAIN_ON_CANCELLATION)

    //需要设置flink checkpoint保存的位置
    env.setStateBackend(new HashMapStateBackend())

    //将状态保存到hdfs中
    env.getCheckpointConfig.setCheckpointStorage("hdfs://master:9000/flink/checkpoint")

    /**
     * 消费kafka中的数据
     *
     */


    val source: KafkaSource[String] = KafkaSource.builder[String]
      .setBootstrapServers("master:9092,node1:9092,node2:9092")
      .setTopics("words")
      .setGroupId("Demo16ExactlyOnce")
      .setStartingOffsets(OffsetsInitializer.earliest) //只在第一次启动的时候生效，如果开启了checkpoint,任务重启之后会按照checkpoint中保证的偏移量消费数据
      .setValueOnlyDeserializer(new SimpleStringSchema())
      .build

    val kafkaSource: DataStream[String] = env.fromSource(source, WatermarkStrategy.noWatermarks(), "Kafka Source")

    val wordsDS: DataStream[String] = kafkaSource.flatMap(_.split(","))

    val kvDS: DataStream[(String, Int)] = wordsDS.map((_, 1))

    val keyByDS: KeyedStream[(String, Int), String] = kvDS.keyBy(_._1)

    val countDS: DataStream[(String, Int)] = keyByDS.sum(1)

    countDS.print()

    env.execute("Demo16ExactlyOnce")
  }
}

4. 从kafka中读取数据，然后再存到kafka中

重点

DeliveryGuarantee.AT_LEAST_ONCE：至少异常，会有重复数据

DeliveryGuarantee.EXACTLY_ONCE： 唯一一次

读取数据的时候.需要指定：

--isolation-level read_committed : 只读已提交的数据

package com.wt.flink.core
import org.apache.flink.api.common.eventtime.WatermarkStrategy
import org.apache.flink.api.common.serialization.SimpleStringSchema
import org.apache.flink.connector.base.DeliveryGuarantee
import org.apache.flink.connector.kafka.sink.{KafkaRecordSerializationSchema, KafkaSink}
import org.apache.flink.connector.kafka.source.KafkaSource
import org.apache.flink.connector.kafka.source.enumerator.initializer.OffsetsInitializer
import org.apache.flink.runtime.state.hashmap.HashMapStateBackend
import org.apache.flink.streaming.api.CheckpointingMode
import org.apache.flink.streaming.api.environment.CheckpointConfig.ExternalizedCheckpointCleanup
import org.apache.flink.streaming.api.scala._

import java.util.Properties

object Demo17ExactlyOnce {
  def main(args: Array[String]): Unit = {

    val env: StreamExecutionEnvironment = StreamExecutionEnvironment.getExecutionEnvironment

    //开启checkpoint

    // 每 1000ms 开始一次 checkpoint
    env.enableCheckpointing(20000)
    // 高级选项：
    // 设置模式为精确一次 (这是默认值)
    env.getCheckpointConfig.setCheckpointingMode(CheckpointingMode.EXACTLY_ONCE)

    // 确认 checkpoints 之间的时间会进行 500 ms
    env.getCheckpointConfig.setMinPauseBetweenCheckpoints(500)

    // Checkpoint 必须在一分钟内完成，否则就会被抛弃
    env.getCheckpointConfig.setCheckpointTimeout(60000)

    // 允许两个连续的 checkpoint 错误
    env.getCheckpointConfig.setTolerableCheckpointFailureNumber(2)

    // 同一时间只允许一个 checkpoint 进行
    env.getCheckpointConfig.setMaxConcurrentCheckpoints(1)

    // 使用 externalized checkpoints，这样 checkpoint 在作业取消后仍就会被保留
    //RETAIN_ON_CANCELLATION: 当任务取消时保留checkpoint
    env.getCheckpointConfig.setExternalizedCheckpointCleanup(
      ExternalizedCheckpointCleanup.RETAIN_ON_CANCELLATION)

    /**
     * 需要设置flink checkpoint保存状态的位置
     *
     */

    env.setStateBackend(new HashMapStateBackend())
    //将状态保存到hdfs中
    env.getCheckpointConfig.setCheckpointStorage("hdfs://master:9000/flink/checkpoint")


    val source: KafkaSource[String] = KafkaSource.builder[String]
      .setBootstrapServers("master:9092,node1:9092,node2:9092")
      .setTopics("source")
      .setGroupId("Demo16ExactlyOnce")
      .setStartingOffsets(OffsetsInitializer.earliest) //只在第一次启动的时候生效，如果开启了checkpoint,任务重启之后会按照checkpoint中保证的偏移量消费数据
      .setValueOnlyDeserializer(new SimpleStringSchema())
      .build

    val kafkaSource: DataStream[String] = env.fromSource(source, WatermarkStrategy.noWatermarks(), "Kafka Source")

    //过滤空数据
    val filterDS: DataStream[String] = kafkaSource.filter(_.nonEmpty)

    /**
     * DeliveryGuarantee.AT_LEAST_ONCE：至少异常，会有重复数据
     * DeliveryGuarantee.EXACTLY_ONCE： 唯一一次
     *
     */

    //将清洗之后的数据保存到kafka中
    val properties = new Properties()
    //设置事务的超时时间，要比15分钟小
    properties.setProperty("transaction.timeout.ms", 10 * 60 * 1000 + "")
    val kafkaSink: KafkaSink[String] = KafkaSink
      .builder[String]()
      .setBootstrapServers("master:9092,node1:9092,node2:9092") //broker地址
      .setKafkaProducerConfig(properties) //设置额外的参数
      .setRecordSerializer(
        KafkaRecordSerializationSchema
          .builder[String]()
          .setTopic("sink") //topic
          .setValueSerializationSchema(new SimpleStringSchema())
          .build())
      .setDeliverGuarantee(DeliveryGuarantee.EXACTLY_ONCE)
      .build()
    
    filterDS.sinkTo(kafkaSink)

    /**
     * 通过命令好消费sink数据
     * --isolation-level read_committed : 只读已提交的数据
     * kafka-console-consumer.sh --bootstrap-server master:9092,node1:9092,node2:9092 --isolation-level read_committed --from-beginning --topic sink
     *
     */
    env.execute()
  }
}