Flink从入门到实践（三）：数据实时采集

文章目录

系列文章索引
一、概述
- 1、版本匹配
- 2、导包
二、编码实现
- 1、基本使用
- 2、更多配置
- 3、自定义序列化器
- 4、Flink SQL方式
三、踩坑
- 1、The MySQL server has a timezone offset (0 seconds ahead of UTC) which does not match the configured timezone Asia/Shanghai.
参考资料

系列文章索引

Flink从入门到实践（一）：Flink入门、Flink部署
Flink从入门到实践（二）：Flink DataStream API
Flink从入门到实践（三）：数据实时采集 - Flink MySQL CDC

一、概述

1、版本匹配

注意MySQL的版本，本次是使用MySQL8.0进行演示。
同时，Flink支持很多数据库的cdc。
在这里插入图片描述
同时也要对应好版本，我们本次使用Flink是1.18，同时FlinkCDC也是3.0版本

2、导包

<dependency><groupId>org.apache.flink</groupId><artifactId>flink-clients</artifactId><version>1.18.0</version>
</dependency>
<dependency><groupId>org.apache.flink</groupId><artifactId>flink-streaming-java</artifactId><version>1.18.0</version>
</dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-connector-base</artifactId><version>1.18.0</version>
</dependency><dependency><groupId>com.ververica</groupId><artifactId>flink-connector-mysql-cdc</artifactId><version>3.0.0</version>
</dependency><dependency><groupId>mysql</groupId><artifactId>mysql-connector-java</artifactId><version>8.0.27</version>
</dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-table-runtime</artifactId><version>1.18.0</version>
</dependency>

二、编码实现

1、基本使用

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;/*** Flink MySql CDC* 每次启动之后，会将所有数据采集一遍*/
public class FlinkCDC01 {public static void main(String[] args) throws Exception {MySqlSource<String> mySqlSource = MySqlSource.<String>builder().hostname("192.168.56.10").port(3306).databaseList("testdb") // 要监听的数据库，可以填多个，支持正则表达式.tableList("testdb.access") // 监听的表，可以填多个，需要db.表，支持正则表达式.username("root").password("root").deserializer(new JsonDebeziumDeserializationSchema()) // converts SourceRecord to JSON String.build();StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();// 开启检查点env.enableCheckpointing(3000);env.fromSource(mySqlSource, WatermarkStrategy.noWatermarks(), "MySQL Source")// 1个并行任务.setParallelism(1).print().setParallelism(1); // 对接收器使用并行性1来保持消息顺序env.execute("Print MySQL Snapshot + Binlog");}
}

结果是json数据：

{
“before”: null,
“after”: {
“id”: 1,
“name”: “1”
},
“source”: {
“version”: “1.9.7.Final”,
“connector”: “mysql”,
“name”: “mysql_binlog_source”,
“ts_ms”: 1707353812000,
“snapshot”: “false”,
“db”: “testdb”, // 库名
“sequence”: null,
“table”: “access”, // 表名
“server_id”: 1,
“gtid”: null,
“file”: “binlog.000005”,
“pos”: 374,
“row”: 0,
“thread”: 9,
“query”: null
},
“op”: “c”, // 操作 c是create；u是update；d是delete；r是read
“ts_ms”: 1707353812450,
“transaction”: null
}

2、更多配置

https://ververica.github.io/flink-cdc-connectors/master/content/connectors/mysql-cdc%28ZH%29.html

配置选项scan.startup.mode指定 MySQL CDC 使用者的启动模式。有效枚举包括：
initial （默认）：在第一次启动时对受监视的数据库表执行初始快照，并继续读取最新的 binlog。
earliest-offset：跳过快照阶段，从可读取的最早 binlog 位点开始读取
latest-offset：首次启动时，从不对受监视的数据库表执行快照，连接器仅从 binlog 的结尾处开始读取，这意味着连接器只能读取在连接器启动之后的数据更改。
specific-offset：跳过快照阶段，从指定的 binlog 位点开始读取。位点可通过 binlog 文件名和位置指定，或者在 GTID 在集群上启用时通过 GTID 集合指定。
timestamp：跳过快照阶段，从指定的时间戳开始读取 binlog 事件。

3、自定义序列化器

import com.ververica.cdc.debezium.DebeziumDeserializationSchema;
import io.debezium.data.Envelope;
import org.apache.flink.api.common.typeinfo.BasicTypeInfo;
import org.apache.flink.api.common.typeinfo.TypeInformation;
import org.apache.flink.util.Collector;
import org.apache.kafka.connect.data.Field;
import org.apache.kafka.connect.data.Struct;
import org.apache.kafka.connect.source.SourceRecord;import java.util.List;public class DomainDeserializationSchema implements DebeziumDeserializationSchema<String> {@Overridepublic void deserialize(SourceRecord sourceRecord, Collector<String> collector) throws Exception {String topic = sourceRecord.topic();String[] split = topic.split("\\.");System.out.println("数据库：" + split[1]);System.out.println("表：" + split[2]);Struct value = (Struct)sourceRecord.value();// 获取before信息Struct before = value.getStruct("before");System.out.println("before：" + before);if (before != null) {// 所有字段List<Field> fields = before.schema().fields();for (Field field : fields) {System.out.println("before field：" + field.name() + " value：" + before.get(field));}}// 获取after信息Struct after = value.getStruct("after");System.out.println("after：" + after);if (after != null) {// 所有字段List<Field> fields = after.schema().fields();for (Field field : fields) {System.out.println("after field：" + field.name() + " value：" + after.get(field));}}// 操作类型Envelope.Operation operation = Envelope.operationFor(sourceRecord);System.out.println("操作：" + operation);// 收集序列化后的结果collector.collect("aaaaaaaaaaaaa");}@Overridepublic TypeInformation<String> getProducedType() {return BasicTypeInfo.STRING_TYPE_INFO; // 类型}
}

MySqlSource<String> mySqlSource = MySqlSource.<String>builder().hostname("192.168.56.10").port(3306).databaseList("testdb") // 要监听的数据库，可以填多个.tableList("testdb.access") // 监听的表，可以填多个.username("root").password("root").deserializer(new DomainDeserializationSchema()) // 序列化器.build();

4、Flink SQL方式

CDC用的少，还是StreamAPI用的多。

三、踩坑

1、The MySQL server has a timezone offset (0 seconds ahead of UTC) which does not match the configured timezone Asia/Shanghai.

2024-02-08 08:36:33 INFO 5217 — [lt-dispatcher-6] o.a.f.r.executiongraph.ExecutionGraph : Source: MySQL Source -> Sink: Print to Std. Out (1/1) (e2371dabd0c952a5dfa7c053cbde80c3_cbc357ccb763df2852fee8c4fc7d55f2_0_2) switched from CREATED to SCHEDULED.
2024-02-08 08:36:33 INFO 5217 — [lt-dispatcher-8] o.a.f.r.r.s.FineGrainedSlotManager : Received resource requirements from job 369b1c979674a0444f679dd13264ea88: [ResourceRequirement{resourceProfile=ResourceProfile{UNKNOWN}, numberOfRequiredSlots=1}]
2024-02-08 08:36:33 INFO 5218 — [lt-dispatcher-6] o.a.flink.runtime.jobmaster.JobMaster : Trying to recover from a global failure.
org.apache.flink.util.FlinkException: Global failure triggered by OperatorCoordinator for ‘Source: MySQL Source -> Sink: Print to Std. Out’ (operator cbc357ccb763df2852fee8c4fc7d55f2).
at org.apache.flink.runtime.operators.coordination.OperatorCoordinatorHolder $L a zy I ni t ia l i ze d C oor d ina t or C o n t e x t . f ai l J o b (Op er a t or C oor d ina t orHo l d er . ja v a : 624) a t or g . a p a c h e . f l ink . r u n t im e . o p er a t ors . coor d ina t i o n . R ecre a t e O n R ese tOp er a t or C oor d ina t or$ QuiesceableContext.failJob(RecreateOnResetOperatorCoordinator.java:248)
at org.apache.flink.runtime.source.coordinator.SourceCoordinatorContext.failJob(SourceCoordinatorContext.java:395)
at org.apache.flink.runtime.source.coordinator.SourceCoordinator.start(SourceCoordinator.java:225)
at org.apache.flink.runtime.operators.coordination.RecreateOnResetOperatorCoordinator $De f err ab l e C oor d ina t or . rese t A n d St a r t (R ecre a t e O n R ese tOp er a t or C oor d ina t or . ja v a : 416) a t or g . a p a c h e . f l ink . r u n t im e . o p er a t ors . coor d ina t i o n . R ecre a t e O n R ese tOp er a t or C oor d ina t or . l amb d a$ resetToCheckpointnull $2 (De f a u ltS c h e d u l er . ja v a : 379) a t ja v a . u t i l . co n c u rre n t . C o m pl e t ab l e F u t u re . u ni R u n (C o m pl e t ab l e F u t u re . ja v a : 719) a t ja v a . u t i l . co n c u rre n t . C o m pl e t ab l e F u t u re$ UniRun.tryFire(CompletableFuture.java:701)
at java.util.concurrent.CompletableFuture $C o m pl e t i o n . r u n (C o m pl e t ab l e F u t u re . ja v a : 456) a t or g . a p a c h e . f l ink . r u n t im e . r p c . p e kk o . P e kk o Rp c A c t or . l amb d a$ handleRunAsync $4 (P e kk o Rp c A c t or . ja v a : 451) a t or g . a p a c h e . f l ink . r u n t im e . co n c u rre n t . Cl a ss L o a d in gU t i l s . r u nWi t h C o n t e x tCl a ss L o a d er (Cl a ss L o a d in gU t i l s . ja v a : 68) a t or g . a p a c h e . f l ink . r u n t im e . r p c . p e kk o . P e kk o Rp c A c t or . han d l e R u n A sy n c (P e kk o Rp c A c t or . ja v a : 451) a t or g . a p a c h e . f l ink . r u n t im e . r p c . p e kk o . P e kk o Rp c A c t or . han d l e Rp c M ess a g e (P e kk o Rp c A c t or . ja v a : 218) a t or g . a p a c h e . f l ink . r u n t im e . r p c . p e kk o . F e n ce d P e kk o Rp c A c t or . han d l e Rp c M ess a g e (F e n ce d P e kk o Rp c A c t or . ja v a : 85) a t or g . a p a c h e . f l ink . r u n t im e . r p c . p e kk o . P e kk o Rp c A c t or . han d l e M ess a g e (P e kk o Rp c A c t or . ja v a : 168) a t or g . a p a c h e . p e kk o . ja p i . p f . U ni tC a se St a t e m e n t . a ppl y (C a se St a t e m e n t s . sc a l a : 33) a t or g . a p a c h e . p e kk o . ja p i . p f . U ni tC a se St a t e m e n t . a ppl y (C a se St a t e m e n t s . sc a l a : 29) a t sc a l a . P a r t ia lF u n c t i o n . a ppl y O r El se (P a r t ia lF u n c t i o n . sc a l a : 127) a t sc a l a . P a r t ia lF u n c t i o n . a ppl y O r El se$ (PartialFunction.scala:126)
at org.apache.pekko.japi.pf.UnitCaseStatement.applyOrElse(CaseStatements.scala:29)
at scala.PartialFunction $O r El se . a ppl y O r El se (P a r t ia lF u n c t i o n . sc a l a : 175) a t sc a l a . P a r t ia lF u n c t i o n$ OrElse.applyOrElse(PartialFunction.scala:176)
at scala.PartialFunction $O r El se . a ppl y O r El se (P a r t ia lF u n c t i o n . sc a l a : 176) a t or g . a p a c h e . p e kk o . a c t or . A c t or . a ro u n d R ece i v e (A c t or . sc a l a : 547) a t or g . a p a c h e . p e kk o . a c t or . A c t or . a ro u n d R ece i v e$ (Actor.scala:545)
at org.apache.pekko.actor.AbstractActor.aroundReceive(AbstractActor.scala:229)
at org.apache.pekko.actor.ActorCell.receiveMessage(ActorCell.scala:590)
at org.apache.pekko.actor.ActorCell.invoke(ActorCell.scala:557)
at org.apache.pekko.dispatch.Mailbox.processMailbox(Mailbox.scala:280)
at org.apache.pekko.dispatch.Mailbox.run(Mailbox.scala:241)
at org.apache.pekko.dispatch.Mailbox.exec(Mailbox.scala:253)
at java.util.concurrent.ForkJoinTask.doExec(ForkJoinTask.java:289)
at java.util.concurrent.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:1067)
at java.util.concurrent.ForkJoinPool.runWorker(ForkJoinPool.java:1703)
at java.util.concurrent.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:172)
Caused by: org.apache.flink.table.api.ValidationException: The MySQL server has a timezone offset (0 seconds ahead of UTC) which does not match the configured timezone Asia/Shanghai. Specify the right server-time-zone to avoid inconsistencies for time-related fields.
at com.ververica.cdc.connectors.mysql.MySqlValidator.checkTimeZone(MySqlValidator.java:184)
at com.ververica.cdc.connectors.mysql.MySqlValidator.validate(MySqlValidator.java:72)
at com.ververica.cdc.connectors.mysql.source.MySqlSource.createEnumerator(MySqlSource.java:197)
at org.apache.flink.runtime.source.coordinator.SourceCoordinator.start(SourceCoordinator.java:221)
… 42 common frames omitted

查看mysql：
show variables like ‘%time_zone%’;
在这里插入图片描述
解决方案：

SET time_zone = 'Asia/Shanghai';
SET @@global.time_zone = 'Asia/Shanghai';
#再次查看
SELECT @@global.time_zone;
show variables like '%time_zone%';

在这里插入图片描述

参考资料

源码：https://github.com/ververica/flink-cdc-connectors
文档：https://ververica.github.io/flink-cdc-connectors/master/content/overview/cdc-connectors.html
官网：https://ververica.github.io/flink-cdc-connectors/