[Flink/FlinkCDC] 實踐總結：Flink 1.12.6 升級 Flink 1.15.4

千千寰宇發表於2024-10-31

Flink DataStream/API

未變的重要特性

雖然官宣建議棄用 JDK 8，使用JDK 11+；但：仍繼續支援 JDK 8

個人猜測：JDK 8 的使用者群實在太大，牽一髮而動全身，防止步子扯太大，遏制自身專案的發展勢頭。

依賴模組的變化

版本變化

  • flink.version : 1.12.6 => 1.15.4
  • flink.connector.version : 1.12.6 => 1.15.4
  • flink.connector.cdc.version : 1.3.0 => 2.3.0

Flink Cdc : flink cdc 2.0.0 之後，【groupId、包路徑】 從 com.alibaba.ververica 變為 com.ververica

  • apache flink cdc 1.3.0
<dependency>
	<groupId>com.alibaba.ververica</groupId>
	<artifactId>flink-connector-mysql-cdc</artifactId>
	<version>1.3.0</version>
</dependency>
  • apache flink cdc 2.3.0
<dependency>
	<groupId>com.alibaba.ververica</groupId>
	<artifactId>flink-connector-mysql-cdc</artifactId>
	<version>2.3.0</version>
</dependency>
  • 詳情參見：
  • Flink CDC 官網： Flink CDC 包 && Flink && JDK && MYSQL 的版本對照 - 部落格園/千千寰宇

各模組擺脫了 scala

詳情參見：

https://github.com/apache/flink/blob/release-1.15.4/docs/content.zh/release-notes/flink-1.15.md 【推薦】
https://nightlies.apache.org/flink/flink-docs-release-1.15/release-notes/flink-1.15/

  • org.apache.flink:flink-clients:${flink.version}

  • flink-streaming-java:

  • org.apache.flink:flink-table-api-java-bridge

org.apache.flink:flink-table-api-java-bridge_${scala.version}:${flink.version}

  • org.apache.flink:flink-connector-kafka:${flink.version}

  • org.apache.flink:flink-runtime-web:${flink.version}

  • `org.apache.flink:flink-statebackend-rocksdb:${flink.version}``

  • org.apache.flink:flink-table-planner:${flink.version}

org.apache.flink:flink-table-planner-blink_${scala.version}:${flink.version}

table-*-blink 轉正 : flink-table-planner/runtime-blink => flink-table-planner、flink-table-runtime

  • 從 Flink 1.15 開始，發行版包含兩個規劃器：
  • flink-table-planner_2.12-${flink.version}.jar : in /opt, 包含查詢規劃器
  • flink-table-planner-loader-${flink.version}.jar【推薦】 : 預設載入/lib，包含隱藏在隔離類路徑後面的查詢計劃器

注意：這2個規劃器(planner_2)不能同時存在於類路徑中。如果將它們都載入到/lib表作業中，則會失敗，報錯Could not instantiate the executor. Make sure a planner module is on the classpath。

Exception in thread "main" org.apache.flink.table.api.TableException: Could not instantiate the executor. Make sure a planner module is on the classpath
    at org.apache.flink.table.api.bridge.internal.AbstractStreamTableEnvironmentImpl.lookupExecutor(AbstractStreamTableEnvironmentImpl.java:108)
    at org.apache.flink.table.api.bridge.java.internal.StreamTableEnvironmentImpl.create(StreamTableEnvironmentImpl.java:100)
    at org.apache.flink.table.api.bridge.java.StreamTableEnvironment.create(StreamTableEnvironment.java:122)
    at org.apache.flink.table.api.bridge.java.StreamTableEnvironment.create(StreamTableEnvironment.java:94)
    at table.FlinkTableTest.main(FlinkTableTest.java:15)
Caused by: org.apache.flink.table.api.ValidationException: Multiple factories for identifier 'default' that implement 'org.apache.flink.table.delegation.ExecutorFactory' found in the classpath.

Ambiguous factory classes are:

org.apache.flink.table.planner.delegation.DefaultExecutorFactory
org.apache.flink.table.planner.loader.DelegateExecutorFactory
    at org.apache.flink.table.factories.FactoryUtil.discoverFactory(FactoryUtil.java:553)
    at org.apache.flink.table.api.bridge.internal.AbstractStreamTableEnvironmentImpl.lookupExecutor(AbstractStreamTableEnvironmentImpl.java:105)
    ... 4 more

Process finished with exit code 1
  • flink 1.14 版本以後，之前版本 flink-table-*-blink-* 轉正。所以：
  • flink-table-planner-blink => flink-table-planner
  • flink-table-runtime-blink => flink-table-runtime

停止支援 scala 2.11，但支援 2.12

scala.version = 2.12
flinkversion = 1.15.4

  • org.apache.flink:flink-connector-hive_${scala.version}:${flink.version}

  • org.apache.flink:flink-table-api-java-bridge_${scala.version}:${flink.version}

相比 flink 1.12.6 時：org.apache.flink:flink-table-api-java-bridge_${scala.version=2.11}:${flink.version=1.12.6}

  • 若報下列錯誤，即：版本不同引起的包衝突。

NoClassDefFoundError: org/apache/flink/shaded/guava30/com/google/common/collect/Lists

原因： flink 1.16、1.15 、1.12.6 等版本使用的 flink-shaded-guava 版本基本不一樣，且版本不相容，需要修改 cdc 中的 flink-shaded-guava 版本。

  • 不同flink版本對應flink-shaded-guava模組的版本
  • flink 1.12.6 : flink-shaded-guava 18.0-12.0
  • flink 1.15.4 : flink-shaded-guava 30.1.1-jre-15.0
  • flink 1.16.0 : flink-shaded-guava 30.1.1-jre-16.0

  • 如果工程內沒有主動引入org.apache.flink:flink-shaded-guava工程，則無需關心此問題————flink-core/flink-runtime/flink-clients等模組內部會預設引入正確的版本

flink 1.15.4

flink 1.12.6

MySQL JDBC Version : ≥ 8.0.16 => ≥8.0.27

  • 版本依據: Apache Flink CDC 官網
  • https://github.com/apache/flink-cdc/tree/release-1.3.0 | ≥8.0.16
  • https://github.com/apache/flink-cdc/tree/release-2.3.0 | ≥8.0.27

針對報錯：Caused by: java.lang.NoSuchMethodError: com.mysql.cj.CharsetMapping.getJavaEncodingForMysqlCharset(Ljava/lang/String;)Ljava/lang/String;

如果MySQL是8.0，fink cdc 2.1 之後由debezium聯結器引起的問題。

  • 將依賴改為8.0.21之後:
<dependency>
	<groupId>mysql</groupId>
	<artifactId>mysql-connector-java</artifactId>
	<version>8.0.32</version>
</dependency>

應用程式的原始碼調整

KafkaRecordDeserializer : 不再存在/不再被支援(flink1.13.0及之後)，並替換為 KafkaDeserializationSchema，KafkaSourceBuilder建立本物件的語法稍有變化

  • org.apache.flink.connector.kafka.source.reader.deserializer.KafkaRecordDeserializer | flink-connector-kafka_2.11 : 1.12.6
  • flink 1.12.6
    https://github.com/apache/flink/blob/release-1.12.6/flink-connectors/flink-connector-kafka/src/main/java/org/apache/flink/connector/kafka/source/reader/deserializer/KafkaRecordDeserializer.java
  • flink 1.12.7 : 仍存在/支援 KafkaRecordDeserializer

https://github.com/apache/flink/blob/release-1.12.7/flink-connectors/flink-connector-kafka/src/main/java/org/apache/flink/connector/kafka/source/reader/deserializer/KafkaRecordDeserializer.java

  • flink 1.13.0 : 不再存在/不再支援 KafkaRecordDeserializer

https://github.com/apache/flink/tree/release-1.13.0/flink-connectors/flink-connector-kafka/src/main/java/org/apache/flink/connector/kafka/source/reader/deserializer

  • flink 14.0

https://github.com/apache/flink/tree/release-1.14.0/flink-connectors/flink-connector-kafka/src/main/java/org/apache/flink/connector/kafka/source/reader/deserializer

  • flink 1.15.4

https://github.com/apache/flink/tree/release-1.15.4/flink-connectors/flink-connector-kafka/src/main/java/org/apache/flink/connector/kafka/source/reader/deserializer/KafkaRecordDeserializationSchema.java

  • flink-connector-kafka : 3.0.0 | 瞭解即可，暫無需被此工程干擾上面思路

https://github.com/apache/flink-connector-kafka/blob/v3.0.0/flink-connector-kafka/src/main/java/org/apache/flink/connector/kafka/source/reader/deserializer/KafkaRecordDeserializationSchema.java

  • 改造原因、改造思路

在 Apache Flink 1.13.0起，KafkaRecordDeserializer已被棄用、並被移除。
如果你正在使用的是Flink的舊版本，並且你看到了KafkaRecordDeserializer的提示，你應該將其替換為使用KafkaDeserializationSchema【推薦】或KafkaDeserializer。
KafkaDeserializationSchema相比KafkaRecordDeserializer，多了需要強制實現的2個方法：

  • boolean isEndOfStream(T var1) : 預設返回 false 即可
  • T deserialize(ConsumerRecord<byte[], byte[]> var1) : 老方法void deserialize(ConsumerRecord<byte[], byte[]> message, Collector<T> out)內部呼叫的即本方法
// flink 1.15.4
//org.apache.flink.streaming.connectors.kafka.KafkaDeserializationSchema

package org.apache.flink.streaming.connectors.kafka;

import java.io.Serializable;
import org.apache.flink.annotation.PublicEvolving;
import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.api.java.typeutils.ResultTypeQueryable;
import org.apache.flink.util.Collector;
import org.apache.kafka.clients.consumer.ConsumerRecord;

@PublicEvolving
public interface KafkaDeserializationSchema<T> extends Serializable, ResultTypeQueryable<T> {
    default void open(DeserializationSchema.InitializationContext context) throws Exception {
    }

    boolean isEndOfStream(T var1);

    T deserialize(ConsumerRecord<byte[], byte[]> var1) throws Exception;//方法1

    default void deserialize(ConsumerRecord<byte[], byte[]> message, Collector<T> out) throws Exception {//方法2
        T deserialized = this.deserialize(message);// 複用/呼叫的方法1
        if (deserialized != null) {
            out.collect(deserialized);
        }
    }
}

故新適配新增的T deserialize(ConsumerRecord<byte[], byte[]> var1)方法是很容易的：

import com.xxx.StringUtils;
import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.api.common.typeinfo.BasicTypeInfo;
import org.apache.flink.api.common.typeinfo.TypeInformation;
import org.apache.flink.api.java.tuple.Tuple2;
import org.apache.flink.api.java.typeutils.TupleTypeInfo;
//import org.apache.flink.connector.kafka.source.reader.deserializer.KafkaRecordDeserializer;
import org.apache.flink.streaming.connectors.kafka.KafkaDeserializationSchema;
import org.apache.flink.util.Collector;
import org.apache.kafka.clients.consumer.ConsumerRecord;

//public class MyKafkaRecordDeserializer implements KafkaRecordDeserializer<Tuple2<String, String>> {
public class MyKafkaRecordDeserializer implements KafkaDeserializationSchema<Tuple2<String, String>> {
/*    @Override
    public void open(DeserializationSchema.InitializationContext context) throws Exception {
        KafkaDeserializationSchema.super.open(context);
    }*/

    @Override
    public boolean isEndOfStream(Tuple2<String, String> stringStringTuple2) {
        return false;
    }

    @Override
    public Tuple2<String, String> deserialize(ConsumerRecord<byte[], byte[]> consumerRecord) throws Exception {//適配新方法1 | 強制
        if(consumerRecord.key() == null){
            return new Tuple2<>("null", StringUtils.bytesToHexString(consumerRecord.value()) );
        }
        return new Tuple2<>( new String(consumerRecord.key() ) , StringUtils.bytesToHexString(consumerRecord.value() ) );
    }

//    @Override
//    public void deserialize(ConsumerRecord<byte[], byte[]> consumerRecord, Collector<Tuple2<String, String>> collector) throws Exception {//適配老方法2 | 非強制
//        collector.collect(new Tuple2<>(consumerRecord.key() == null ? "null" : new String(consumerRecord.key()), StringUtils.bytesToHexString(consumerRecord.value())));
//    }

    @Override
    public TypeInformation<Tuple2<String, String>> getProducedType() {
        return new TupleTypeInfo<>(BasicTypeInfo.STRING_TYPE_INFO, BasicTypeInfo.STRING_TYPE_INFO);
    }
}

使用本類、建立本類物件的方式，也稍有變化：

// org.apache.flink.connector.kafka.source.KafkaSourceBuilder | flink-connector-kafka:1.15.4
KafkaSourceBuilder<Tuple2<String, String>> kafkaConsumerSourceBuilder = KafkaSource.<Tuple2<String, String>>builder()
	.setTopics(canTopic)
	.setProperties(kafkaConsumerProperties)
	.setClientIdPrefix(Constants.JOB_NAME + "#" + System.currentTimeMillis() + "")
	.setDeserializer( KafkaRecordDeserializationSchema.of(new MyKafkaRecordDeserializer()) ); // flink 1.15.4
	//.setDeserializer(new MyKafkaRecordDeserializer());// flink 1.12.6
  • 推薦文獻
  • Flink1.14新版KafkaSource和KafkaSink實踐使用(自定義反序列化器、Topic選擇器、序列化器、分割槽器) - CSDN
  • https://nightlies.apache.org/flink/flink-docs-release-1.15/zh/docs/connectors/datastream/kafka/

Flink Cdc : flink cdc 2.0.0 之後，【groupId、包路徑】 從 com.alibaba.ververica 變為 com.ververica

MySQLSource : 包路徑被調整(2.0.0及之後)、類名大小寫有變化(flink cdc 2.0.0 及之後)、不再被推薦使用(flink cdc 2.1.0 及之後)

  • com.alibaba.ververica.cdc.connectors.mysql.MySQLSource | flink cdc 1.3.0

https://github.com/apache/flink-cdc/blob/release-1.3.0/flink-connector-mysql-cdc/src/main/java/com/alibaba/ververica/cdc/connectors/mysql/MySQLSource.java
包路徑被調整、類名大小寫有變化

https://github.com/apache/flink-cdc/blob/release-2.0.0/flink-connector-mysql-cdc/src/main/java/com/ververica/cdc/connectors/mysql/MySqlSource.java
com.ververica.cdc.connectors.mysql.MySqlSource 自 flink cdc 2.1.0 及之後被建議棄用、但com.ververica.cdc.connectors.mysql.source.MySqlSource被推薦可用
https://github.com/apache/flink-cdc/blob/release-2.1.0/flink-connector-mysql-cdc/src/main/java/com/ververica/cdc/connectors/mysql/MySqlSource.java
Flink CDC這個MySqlSource棄用了，還有別的方式嗎？ - aliyun 【推薦】

有兩個MysqlSource，一個是棄用的，另一個是可用的，包名不同。com.ververica.cdc.connectors.mysql.source這個包下的是可用的。

  • com.ververica.cdc.connectors.mysql.source.MySqlSource | flink cdc 2.3.0

https://github.com/apache/flink-cdc/blob/release-2.3.0/flink-connector-mysql-cdc/src/main/java/com/ververica/cdc/connectors/mysql/MySqlSource.java

serverId : 如果選擇新的MySqlSource類，則：其設定入參稍有變化
  • com.alibaba.ververica.cdc.connectors.mysql.MySQLSource#serverId() | flink cdc 1.3.0

https://github.com/apache/flink-cdc/blob/release-1.3.0/flink-connector-mysql-cdc/src/main/java/com/alibaba/ververica/cdc/connectors/mysql/MySQLSource.java

  • com.ververica.cdc.connectors.mysql.source.MySqlSource | flink cdc 2.1.0 、 2.3.0 【被推薦使用】

https://github.com/apache/flink-cdc/blob/release-2.1.0/flink-connector-mysql-cdc/src/main/java/com/ververica/cdc/connectors/mysql/source/MySqlSource.java

沒有serverId方法
https://github.com/apache/flink-cdc/blob/release-2.1.0/flink-connector-mysql-cdc/src/main/java/com/ververica/cdc/connectors/mysql/source/MySqlSourceBuilder.java
有serverId方法，透過MySqlSource.<String>builder()即MySqlSourceBuilder

/**
 * A numeric ID or a numeric ID range of this database client, The numeric ID syntax is like
 * '5400', the numeric ID range syntax is like '5400-5408', The numeric ID range syntax is
 * required when 'scan.incremental.snapshot.enabled' enabled. Every ID must be unique across all
 * currently-running database processes in the MySQL cluster. This connector joins the MySQL
 * cluster as another server (with this unique ID) so it can read the binlog. By default, a
 * random number is generated between 5400 and 6400, though we recommend setting an explicit
 * value."
 */
public MySqlSourceBuilder<T> serverId(String serverId) {
	this.configFactory.serverId(serverId);
	return this;
}
  • com.ververica.cdc.connectors.mysql.source.MySqlSource#serverId(int serverId) | flink cdc 2.1.0 【被建議棄用】、flink cdc 2.3.0 【被廢止/無法用】

https://github.com/apache/flink-cdc/blob/release-2.1.0/flink-connector-mysql-cdc/src/main/java/com/ververica/cdc/connectors/mysql/MySqlSource.java

/**
 * A numeric ID of this database client, which must be unique across all currently-running
 * database processes in the MySQL cluster. This connector joins the MySQL database cluster
 * as another server (with this unique ID) so it can read the binlog. By default, a random
 * number is generated between 5400 and 6400, though we recommend setting an explicit value.
 */
public Builder<T> serverId(int serverId) {
	this.serverId = serverId;
	return this;
}
  • 改造Demo: flink 1.3.0
SourceFunction<String> mySqlSource = 
	MySqlSource.<String>builder()
	//資料庫地址
	.hostname(jobParameterTool.get("cdc.mysql.hostname"))
	//埠號
	.port(Integer.parseInt(jobParameterTool.get("cdc.mysql.port")))
	//使用者名稱
	.username(jobParameterTool.get("cdc.mysql.username"))
	//密碼
	.password(jobParameterTool.get("cdc.mysql.password"))
	//監控的資料庫
	.databaseList(jobParameterTool.get("cdc.mysql.databaseList"))
	//監控的表名，格式資料庫.表名
	.tableList(jobParameterTool.get("cdc.mysql.tableList"))
	//虛擬化方式
	.deserializer(new MySQLCdcMessageDeserializationSchema())
	//時區
	.serverTimeZone("UTC")
	.serverId( randomServerId(5000, Constants.JOB_NAME + "#xxxConfig") )
	.startupOptions(StartupOptions.latest())
	.build();


public static Integer randomServerId(int interval, String jobCdcConfigDescription){
	//startServerId ∈[ interval + 0, interval + interval)
	//int serverId = RANDOM.nextInt(interval) + interval; // RANDOM.nextInt(n) : 生成介於 [0,n) 區間的隨機整數
	//serverId = [ 7000 + 0, Integer.MAX_VALUE - interval)
	int serverId = RANDOM.nextInt(Integer.MAX_VALUE - interval - 7000) + 7000;
	log.info("Success to generate random server id result! serverId : {}, interval : {}, jobCdcConfigDescription : {}"
			, serverId , interval , jobCdcConfigDescription );
	return serverId;
}
  • 改造Demo: flink 2.3.0
MySqlSource<String> mySqlSource = 
	MySqlSource.<String>builder()
	//資料庫地址
	.hostname(jobParameterTool.get("cdc.mysql.hostname"))
	//埠號
	.port(Integer.parseInt(jobParameterTool.get("cdc.mysql.port")))
	//使用者名稱
	.username(jobParameterTool.get("cdc.mysql.username"))
	//密碼
	.password(jobParameterTool.get("cdc.mysql.password"))
	//監控的資料庫
	.databaseList(jobParameterTool.get("cdc.mysql.databaseList"))
	//監控的表名，格式資料庫.表名
	.tableList(jobParameterTool.get("cdc.mysql.tableList"))
	//虛擬化方式
	.deserializer(new MySQLCdcMessageDeserializationSchema())
	//時區
	.serverTimeZone("UTC")
	.serverId( randomServerIdRange(5000, Constants.JOB_NAME + "#xxxConfig") )
	.startupOptions(StartupOptions.latest())
	.build();


//新增強制要求: interval >= 本運算元的並行度
public static String randomServerIdRange(int interval, String jobCdcConfigDescription){
	// 生成1個起始隨機數 |
	//startServerId = [interval + 0, interval + interval )
	//int startServerId = RANDOM.nextInt(interval) + interval; // RANDOM.nextInt(n) : 生成介於 [0,n) 區間的隨機整數
	//startServerId = [ 7000 + 0, Integer.MAX_VALUE - interval)
	int startServerId = RANDOM.nextInt(Integer.MAX_VALUE - interval - 7000) + 7000;

	//endServerId ∈ [startServerId, startServerId + interval];
	int endServerId = startServerId + interval;
	log.info("Success to generate random server id result! startServerId : {},endServerId : {}, interval : {}, jobCdcConfigDescription : {}"
			, startServerId, endServerId , interval , jobCdcConfigDescription );
	return String.format("%d-%d", startServerId, endServerId);
}
MySQLSourceBuilder#build 方法: 返回型別存在變化: SourceFunction/DebeziumSourceFunction<T> => MySqlSource<T>
  • org.apache.flink.streaming.api.functions.source.SourceFunction => com.ververica.cdc.connectors.mysql.source.MySqlSource
//com.alibaba.ververica.cdc.connectors.mysql.MySQLSource.Builder#build | flink cdc 1.3.0
// 返回: com.alibaba.ververica.cdc.debezium.DebeziumSourceFunction
// public class DebeziumSourceFunction<T> extends RichSourceFunction<T> implements CheckpointedFunction, CheckpointListener, ResultTypeQueryable<T>
//public abstract class org.apache.flink.streaming.api.functions.source.RichSourceFunction<OUT> extends AbstractRichFunction implements SourceFunction<OUT>
public DebeziumSourceFunction<T> build() {
	Properties props = new Properties();
	props.setProperty("connector.class", MySqlConnector.class.getCanonicalName());
	props.setProperty("database.server.name", "mysql_binlog_source");
	props.setProperty("database.hostname", (String)Preconditions.checkNotNull(this.hostname));
	props.setProperty("database.user", (String)Preconditions.checkNotNull(this.username));
	props.setProperty("database.password", (String)Preconditions.checkNotNull(this.password));
	props.setProperty("database.port", String.valueOf(this.port));
	props.setProperty("database.history.skip.unparseable.ddl", String.valueOf(true));
	if (this.serverId != null) {
		props.setProperty("database.server.id", String.valueOf(this.serverId));
	}
	...
}


//com.ververica.cdc.connectors.mysql.source.MySqlSourceBuilder#build | flink cdc 2.3.0
//// 返回: 
public MySqlSource<T> build() {
	return new MySqlSource(this.configFactory, (DebeziumDeserializationSchema)Preconditions.checkNotNull(this.deserializer));
}
  • 使用變化Demo: Flink cdc 1.3.0

mysqlSource 想要監聽 mysql 表結構變更（例如：新增新的欄位），要怎麼辦？設定 - aliyun

Properties properties = new Properties();
properties.setProperty("database.hostname", "localhost");
properties.setProperty("database.port", "3306");
properties.setProperty("database.user", "your_username");
properties.setProperty("database.password", "your_password");
properties.setProperty("database.server.id", "1"); // 設定唯一的 server id
properties.setProperty("database.server.name", "mysql_source");

DebeziumSourceFunction<String> sourceFunction = MySQLSource.<String>builder()
    .hostname("localhost")
    .port(3306)
    .username("your_username")
    .password("your_password")
    .databaseList("your_database")
    .tableList("your_table")
    .includeSchemaChanges(true) // 開啟監聽表結構變更
    .deserializer(new StringDebeziumDeserializationSchema())
    .build();

DataStreamSource<String> stream = env.addSource(sourceFunction);//可以使用 addSource

stream.print();
env.execute("MySQL CDC Job");
  • 使用變化Demo: Flink cdc 2.3.0

https://flink-tpc-ds.github.io/flink-cdc-connectors/release-2.3/content/connectors/mysql-cdc(ZH).html
無法使用 env.addSource(SourceFunction, String sourceName)，只能使用env.fromSource(Source<OUT, ?, ?> source, WatermarkStrategy<OUT> timestampsAndWatermarks, String sourceName)

import org.apache.flink.api.common.eventtime.WatermarkStrategy;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import com.ververica.cdc.debezium.JsonDebeziumDeserializationSchema;
import com.ververica.cdc.connectors.mysql.source.MySqlSource;

public class MySqlSourceExample {
  public static void main(String[] args) throws Exception {
    MySqlSource<String> mySqlSource = MySqlSource.<String>builder()
        .hostname("yourHostname")
        .port(yourPort)
        .databaseList("yourDatabaseName") // 設定捕獲的資料庫， 如果需要同步整個資料庫，請將 tableList 設定為 ".*".
        .tableList("yourDatabaseName.yourTableName") // 設定捕獲的表
        .username("yourUsername")
        .password("yourPassword")
        .deserializer(new JsonDebeziumDeserializationSchema()) // 將 SourceRecord 轉換為 JSON 字串
        .build();

    StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();

    // 設定 3s 的 checkpoint 間隔
    env.enableCheckpointing(3000);

    env
      .fromSource(mySqlSource, WatermarkStrategy.noWatermarks(), "MySQL Source")
      // 設定 source 節點的並行度為 4
      .setParallelism(4)
      .print().setParallelism(1); // 設定 sink 節點並行度為 1 

    env.execute("Print MySQL Snapshot + Binlog");
  }
}

StartupOptions : 包路徑被調整(2.0.0及之後)

  • import com.alibaba.ververica.cdc.connectors.mysql.table.StartupOptions | flink 1.3.0

https://github.com/apache/flink-cdc/blob/release-1.3.0/flink-connector-mysql-cdc/src/main/java/com/alibaba/ververica/cdc/connectors/mysql/table/StartupOptions.java

  • com.ververica.cdc.connectors.mysql.table.StartupOptions | flink 2.3.0

https://github.com/apache/flink-cdc/blob/release-2.3.0/flink-connector-mysql-cdc/src/main/java/com/ververica/cdc/connectors/mysql/table/StartupOptions.java

DebeziumDeserializationSchema : 包路徑被調整(flink-cdc2.0.0及之後)

  • com.alibaba.ververica.cdc.debezium.DebeziumDeserializationSchema | flink cdc 1.3.0

com.ververica:flink-connector-debezium:1.3.0
https://github.com/apache/flink-cdc/blob/release-1.3.0/flink-connector-debezium/src/main/java/com/alibaba/ververica/cdc/debezium/DebeziumDeserializationSchema.java

  • com.ververica.cdc.debezium.DebeziumDeserializationSchema | flink cdc 2.3.0

com.ververica:flink-connector-debezium:2.3.0
https://github.com/apache/flink-cdc/blob/release-2.3.0/flink-connector-debezium/src/main/java/com/ververica/cdc/debezium/DebeziumDeserializationSchema.java

X 參考文獻

  • Flink+Flink CDC版本升級的依賴問題總結 - CSDN
  • apache flink cdc
  • https://github.com/apache/flink-cdc
  • https://github.com/apache/flink-cdc/blob/master/docs/content/docs/faq/faq.md
  • https://github.com/apache/flink-cdc/tree/master/flink-cdc-connect/flink-cdc-source-connectors/flink-sql-connector-mysql-cdc
  • com.alibaba.ververica:flink-connector-mysql-cdc:1.3.0

https://github.com/apache/flink-cdc/blob/release-1.3.0/flink-connector-mysql-cdc/pom.xml 【推薦】 Flink 1.12.6

  • com.ververica:flink-connector-mysql-cdc:2.0

MYSQL (Database: 5.7, 8.0.x / JDBC Driver: 8.0.16 ) | Flink 1.12 + | JDK 8+
https://github.com/apache/flink-cdc/tree/release-2.0
https://github.com/apache/flink-cdc/blob/release-2.0/flink-connector-mysql-cdc/pom.xml

  • com.ververica:flink-connector-mysql-cdc:2.3.0

https://github.com/apache/flink-cdc/blob/release-2.3.0/flink-connector-mysql-cdc/pom.xml 【推薦】 Flink 1.15.4

  • org.apache.flink:flink-connector-mysql-cdc:${flink.cdc.version}
  • https://ververica.github.io/flink-cdc-connectors/master/content/about.html [已廢止]
  • apache flink
  • https://github.com/apache/flink
  • https://flink.apache.org
  • apache flink-connector-kafka
  • https://github.com/apache/flink-connector-kafka
  • https://flink.apache.org

相關文章