ToB企服应用市场:ToB评测及商务社交产业平台

标题: Flink CDC 使用实践以及遇到的问题 [打印本页]

作者: 北冰洋以北 时间: 2024-11-28 16:43
标题: Flink CDC 使用实践以及遇到的问题
背景

近来公司在做一些业务上的架构调整，有一部门是数据从mysql采集到Starrocks，之前的一套方法是走 debezium 到 puslar 到 starrocks,这一套下来比较需要设置很多东西，而且出现问题以后，需要修改很多设置，而且现阶段问题比较多，且采集的是全部线上的数据库，维护起来很费劲。
于是我们举行了采集的数据流调整，使用 Flink CDC这一套，这一套是端到端的，且采取设置化的方式，支持schema的变动，无需再多一层中间存储层。
终极设置

关于flink cdc的使用设置可以参考Flink CDC 源码分析–整体流程，我能这里只贴出来我们终极使用的设置：

source:
type: mysql
name: Database mysql to Data warehouse
hostname: xxxx
port: 3306
username: xxx
password: xxx
tables: db1.table1
server-id: 556401-556500
scan.startup.mode: initial
scan.snapshot.fetch.size: 8096
scan.incremental.snapshot.chunk.size: 16192
debezium.max.queue.size: 162580
debezium.max.batch.size: 40960
debezium.poll.interval.ms: 50
scan.only.deserialize.captured.tables.changelog.enabled: true
scan.parallel-deserialize-changelog.enabled: true
heartbeat.interval: 5s
scan.newly-added-table.enabled: true
sink:
type: starrocks
name: StarRocks Sink
jdbc-url: xxx
load-url: xxx
username: xxx
password: xxx
sink.buffer-flush.interval-ms: 5000
table.create.properties.replication_num: 3
table.create.num-buckets: 3
route:
- source-table: db1.\.*
sink-table: db1.<>
replace-symbol: <>
description: route all tables to starrrocks
pipeline:
name: Sync mysql Database to StarRocks
parallelism: 1
schema.change.behavior: EVOLVE

复制代码

遇到的问题

EventHeaderV4反序列化问题
报错如下：
1. Caused by: io.debezium.DebeziumException: Failed to deserialize data of EventHeaderV4{timestamp=1732257303000, eventType=WRITE_ROWS, serverId=28555270, headerLength=19, dataLength=320, nextPosition=383299196, flags=0}
2. at io.debezium.connector.mysql.MySqlStreamingChangeEventSource.wrap(MySqlStreamingChangeEventSource.java:1718)
3. ... 5 more
4. Caused by: com.github.shyiko.mysql.binlog.event.deserialization.EventDataDeserializationException: Failed to deserialize data of EventHeaderV4{timestamp=1732257303000, eventType=WRITE_ROWS, serverId=28555270, headerLength=19, dataLength=320, nextPosition=383299196, flags=0}
5. at com.github.shyiko.mysql.binlog.event.deserialization.EventDeserializer.deserializeEventData(EventDeserializer.java:358)
6. at com.github.shyiko.mysql.binlog.event.deserialization.EventDeserializer.nextEvent(EventDeserializer.java:252)
7. at io.debezium.connector.mysql.MySqlStreamingChangeEventSource$1.nextEvent(MySqlStreamingChangeEventSource.java:388)
8. at com.github.shyiko.mysql.binlog.BinaryLogClient.listenForEventPackets(BinaryLogClient.java:1187)
9. ... 3 more
10. Caused by: java.io.EOFException: Failed to read remaining 28 of 36 bytes from position 258280448. Block length: 183. Initial block length: 316.
11. at com.github.shyiko.mysql.binlog.io.ByteArrayInputStream.fill(ByteArrayInputStream.java:115)
12. at com.github.shyiko.mysql.binlog.io.ByteArrayInputStream.read(ByteArrayInputStream.java:105)
13. at io.debezium.connector.mysql.RowDeserializers.deserializeVarString(RowDeserializers.java:264)
14. at io.debezium.connector.mysql.RowDeserializers$WriteRowsDeserializer.deserializeVarString(RowDeserializers.java:192)
15. at com.github.shyiko.mysql.binlog.event.deserialization.AbstractRowsEventDataDeserializer.deserializeCell(AbstractRowsEventDataDeserializer.java:189)
16. at com.github.shyiko.mysql.binlog.event.deserialization.AbstractRowsEventDataDeserializer.deserializeRow(AbstractRowsEventDataDeserializer.java:143)
17. at com.github.shyiko.mysql.binlog.event.deserialization.WriteRowsEventDataDeserializer.deserializeRows(WriteRowsEventDataDeserializer.java:75)
18. at com.github.shyiko.mysql.binlog.event.deserialization.WriteRowsEventDataDeserializer.deserialize(WriteRowsEventDataDeserializer.java:65)
19. at com.github.shyiko.mysql.binlog.event.deserialization.WriteRowsEventDataDeserializer.deserialize(WriteRowsEventDataDeserializer.java:38)
20. at com.github.shyiko.mysql.binlog.event.deserialization.EventDeserializer.deserializeEventData(EventDeserializer.java:352)
21. ... 6 more
复制代码
过段时间自己恢复
这个征象比较诡异，过段时间就自己恢复了，目前怀疑的点：
- mysql毗连数和带宽问题
- msyql服务端的设置问题，可以参考Flink CDC FAQ
  1. mysql> set global slave_net_timeout = 120;
  2. mysql> set global thread_pool_idle_timeout = 120;
  复制代码
- 作业反压导致，参考阿里云Flink
  1. execution.checkpointing.interval=10min
  2. execution.checkpointing.tolerable-failed-checkpoints=100
  3. debezium.connect.keep.alive.interval.ms = 40000
  复制代码
Starrocks Be 内存受限
1. java.lang.RuntimeException: com.starrocks.data.load.stream.exception.StreamLoadFailException: Transaction prepare failed, db: shoufuyou_fund, table: fund_common_repay_push, label: flink-4c6c8cfb-5116-4c38-a60e-a1b87cd6f2f2,
2. responseBody: {
3. "Status": "MEM_LIMIT_EXCEEDED",
4. "Message": "Memory of process exceed limit. QUERY Backend: 172.17.172.251, fragment: 9048ed6e-6ffb-04db-081b-a4966b179387 Used: 26469550752, Limit: 26316804096. Mem usage has exceed the limit of BE"
5. }
6. errorLog: null
7. at com.starrocks.data.load.stream.v2.StreamLoadManagerV2.AssertNotException(StreamLoadManagerV2.java:427)
8. at com.starrocks.data.load.stream.v2.StreamLoadManagerV2.write(StreamLoadManagerV2.java:252)
9. at com.starrocks.connector.flink.table.sink.v2.StarRocksWriter.write(StarRocksWriter.java:143)
10. at org.apache.flink.streaming.runtime.operators.sink.SinkWriterOperator.processElement(SinkWriterOperator.java:182)
11. at org.apache.flink.cdc.runtime.operators.sink.DataSinkWriterOperator.processElement(DataSinkWriterOperator.java:178)
12. at org.apache.flink.streaming.runtime.tasks.CopyingChainingOutput.pushToOperator(CopyingChainingOutput.java:75)
13. at org.apache.flink.streaming.runtime.tasks.CopyingChainingOutput.collect(CopyingChainingOutput.java:50)
14. at org.apache.flink.streaming.runtime.tasks.CopyingChainingOutput.collect(CopyingChainingOutput.java:29)
15. at org.apache.flink.streaming.api.operators.StreamMap.processElement(StreamMap.java:38)
16. at org.apache.flink.streaming.runtime.tasks.OneInputStreamTask$StreamTaskNetworkOutput.emitRecord(OneInputStreamTask.java:245)
17. at org.apache.flink.streaming.runtime.io.AbstractStreamTaskNetworkInput.processElement(AbstractStreamTaskNetworkInput.java:217)
18. at org.apache.flink.streaming.runtime.io.AbstractStreamTaskNetworkInput.emitNext(AbstractStreamTaskNetworkInput.java:169)
19. at org.apache.flink.streaming.runtime.io.StreamOneInputProcessor.processInput(StreamOneInputProcessor.java:68)
20. at org.apache.flink.streaming.runtime.tasks.StreamTask.processInput(StreamTask.java:616)
21. at org.apache.flink.streaming.runtime.tasks.mailbox.MailboxProcessor.runMailboxLoop(MailboxProcessor.java:231)
22. at org.apache.flink.streaming.runtime.tasks.StreamTask.runMailboxLoop(StreamTask.java:1071)
23. at org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:1020)
24. at org.apache.flink.runtime.taskmanager.Task.runWithSystemExitMonitoring(Task.java:959)
25. at org.apache.flink.runtime.taskmanager.Task.restoreAndInvoke(Task.java:938)
26. at org.apache.flink.runtime.taskmanager.Task.doRun(Task.java:751)
27. at org.apache.flink.runtime.taskmanager.Task.run(Task.java:567)
28. at java.lang.Thread.run(Thread.java:879)
29. Caused by: com.starrocks.data.load.stream.exception.StreamLoadFailException: Transaction prepare failed, db: shoufuyou_fund, table: fund_common_repay_push, label: flink-4c6c8cfb-5116-4c38-a60e-a1b87cd6f2f2,
30. responseBody: {
31. "Status": "MEM_LIMIT_EXCEEDED",
32. "Message": "Memory of process exceed limit. QUERY Backend: 172.17.172.251, fragment: 9048ed6e-6ffb-04db-081b-a4966b179387 Used: 26469550752, Limit: 26316804096. Mem usage has exceed the limit of BE"
33. }
34. errorLog: null
35. at com.starrocks.data.load.stream.TransactionStreamLoader.prepare(TransactionStreamLoader.java:221)
36. at com.starrocks.data.load.stream.v2.TransactionTableRegion.commit(TransactionTableRegion.java:247)
37. at com.starrocks.data.load.stream.v2.StreamLoadManagerV2.lambda$init$0(StreamLoadManagerV2.java:210)
38. ... 1 more
复制代码
由于我们 Starrocks BE的内存是在 32GB，开启多个Flink CDC 使命，会导致CDC初始化的时候，写入BE的数据太多，从而BE内存不敷，
办理：低落写入Starrocks的并行读，不要太多CDC同时并行
也可以参考Troubleshooting StarRocks memory hog issues
JobManager Direct buffer memory不敷
1. java. lang.OutOfMemoryError: Direct buffer memory
2. at lava.n10.B1ts.reserveMemory(B1ts.lava:/08 ~ 7:1.8.0 312.
3. at java.nio.DirectByteBuffer.<init>(DirectByteBuffer.java:123) ~[7:1.8.0_372]
4. at java.nio.ByteBuffer.allocateDirect(ByteBuffer.java:311) ~[7:1.8.0_3721
5. at sun.nio.ch.Util.getTemporaryDirectBuffer(Util.java:247) ~[7:1.8.0_372]
6. at sun.nio.ch.IOUtil.write(IOUtil.java:60) ~[7:1.8.0_372]
7. at sun.nio.ch.FileChannelImpl.write(FileChannelImpl.java:234) ~[?:1.8.0_372]
8. at java.nio.channels.Channels.writeFullyImpl(Channels.java:78) ~[?:1.8.0_372]
9. at java.nio.channels.Channels$1.write(Channels.java:174) ~[7:1.8.0_372]
10. at org.apache.flink.core.fs.OffsetAware0utputStream.write(0ffsetAware0utputStream.java:48) ~[ververica-connector-vvp-1.17-vvr-8.0.9-2-SNAPSHOT-jar-with-dependencies.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
11. at org.apache.flink.core.fs.RefCountedFileWithStream.write(RefCountedFileWithStream.java:54) ~[ververica-connector-vvp-1.17-vvr-8.0.9-2-SNAPSHOT-jar-with-dependencies. jar: 1.17-vvr-8.0.9-2-SNAPSHOT
12. at org.apache.flink.core.fs.RefCountedBufferingFileStream.write(RefCountedBufferingFileStream.java:88) ~[ververica-connector-vvp-1.17-vvr-8.0.9-2-SNAPSHOT-jar-with-dependencies.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
13. at ora.aoache.flink.fs.osshadooo.writer.OSSRecoverableFsDataOutoutStream.write OSSRecoverableFsDataOutoutStream.1ava:130) ~?:?
14. at org.apache.flink. runtime.state.filesystem.FsCheckpointMetadata0utputStream.write(FsCheckpointMetadata0utputStream.java:78) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT. jar:1.17-vvr-8.0.9-2-SNAPSHOT]
15. at java.io.Data0utputStream.write(DataOutputStream.java:107) ~[7:1.8.0_372]
16. at java.io.Filter0utputStream.write(FilterOutputStream.java:97) ~[7:1.8.0_372]
17. at org.apache.flink. runtime.checkpoint.metadata.MetadataV2V3SerializerBase.serializeStreamStateHandle(MetadataV2V3SerializerBase.java:703) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT. jar:1.17-vvr-8.0.9-2-SNAPSHOT]
18. at org.apache. flink.runtime.checkpoint.metadata.MetadataV3Serializer.serializeStreamStateHandle(MetadataV3Serializer.java:264) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT] at org.apache.flink. runtime.checkpoint.metadata.MetadataV3Serializer.serialize0peratorState(MetadataV3Serializer.java:109) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
19. at org.apache. flink. runtime.checkpoint.metadata.MetadataV2V3SerializerBase.serializeMetadata(MetadataV2V3SerializerBase.java:153) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
20. at org.apache.flink. runtime.checkpoint.metadata.MetadataV3Serializer.serialize(MetadataV3Serializer.java:83) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT] at org.apache.flink.runtime.checkpoint.metadata.MetadataV4Serializer.serialize(MetadataV4Serializer.java:56) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]|
21. at org.apache. flink. runtime.checkpoint.Checkpoints.storeCheckpointMetadata(Checkpoints. java:102) ~[flink-dist-1.17-vvr-8.0.
复制代码
办理：
增加设置:
1. jobmanager.memory.off-heap.size: 512mb
复制代码
TaskManager jvm内存不敷
1. java.util.concurrent.TimeoutException: Heartbeat of TaskManager with id job-da2375f5-405b-4398-a568-eaba9711576d-taskmanager-1-34 timed out.
2. at org.apache.flink.runtime.jobmaster.JobMaster$TaskManagerHeartbeatListener.notifyHeartbeatTimeout(JobMaster.java:1714)
3. at org.apache.flink.runtime.heartbeat.DefaultHeartbeatMonitor.run(DefaultHeartbeatMonitor.java:158)
4. at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
5. at java.util.concurrent.FutureTask.run(FutureTask.java:266)
6. at org.apache.flink.runtime.rpc.akka.AkkaRpcActor.lambda$handleRunAsync$4(AkkaRpcActor.java:453)
7. at org.apache.flink.runtime.concurrent.akka.ClassLoadingUtils.runWithContextClassLoader(ClassLoadingUtils.java:68)
8. at org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRunAsync(AkkaRpcActor.java:453)
9. at org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleRpcMessage(AkkaRpcActor.java:218)
10. at org.apache.flink.runtime.rpc.akka.FencedAkkaRpcActor.handleRpcMessage(FencedAkkaRpcActor.java:84)
11. at org.apache.flink.runtime.rpc.akka.AkkaRpcActor.handleMessage(AkkaRpcActor.java:168)
12. at akka.japi.pf.UnitCaseStatement.apply(CaseStatements.scala:24)
13. at akka.japi.pf.UnitCaseStatement.apply(CaseStatements.scala:20)
14. at scala.PartialFunction.applyOrElse(PartialFunction.scala:127)
15. at scala.PartialFunction.applyOrElse$(PartialFunction.scala:126)
16. at akka.japi.pf.UnitCaseStatement.applyOrElse(CaseStatements.scala:20)
17. at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:175)
18. at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:176)
19. at scala.PartialFunction$OrElse.applyOrElse(PartialFunction.scala:176)
20. at akka.actor.Actor.aroundReceive(Actor.scala:537)
21. at akka.actor.Actor.aroundReceive$(Actor.scala:535)
22. at akka.actor.AbstractActor.aroundReceive(AbstractActor.scala:220)
23. at akka.actor.ActorCell.receiveMessage(ActorCell.scala:579)
24. at akka.actor.ActorCell.invoke(ActorCell.scala:547)
25. at akka.dispatch.Mailbox.processMailbox(Mailbox.scala:270)
26. at akka.dispatch.Mailbox.run(Mailbox.scala:231)
27. at akka.dispatch.Mailbox.exec(Mailbox.scala:243)
28. at java.util.concurrent.ForkJoinTask.doExec(ForkJoinTask.java:289)
29. at java.util.concurrent.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:1056)
30. at java.util.concurrent.ForkJoinPool.runWorker(ForkJoinPool.java:1692)
31. at java.util.concurrent.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:175)
复制代码
办理：
在运行的过程中我们发现TaskManager的 taskmanager.memory.managed.size 内存使用一直为0，这是因为我们这里没有状态的存储，只是ETL,可以参考Flink TaskManager Memory Model

所以增加以下设置
1. taskmanager.memory.managed.size: 256mb
2. taskmanager.memory.process.size: 4096m
3. table.exec.state.ttl: 1 m
复制代码
读取mysql数据过慢
1. java.lang.RuntimeException: One or more fetchers have encountered exception
2. at org.apache.flink.connector.base.source.reader.fetcher.SplitFetcherManager.checkErrors(SplitFetcherManager.java:261) ~[flink-connector-files-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
3. at org.apache.flink.connector.base.source.reader.SourceReaderBase.getNextFetch(SourceReaderBase.java:185) ~[flink-connector-files-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
4. at org.apache.flink.connector.base.source.reader.SourceReaderBase.pollNext(SourceReaderBase.java:144) ~[flink-connector-files-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
5. at org.apache.flink.streaming.api.operators.SourceOperator.pollNext(SourceOperator.java:779) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
6. at org.apache.flink.streaming.api.operators.SourceOperator.emitNext(SourceOperator.java:457) ~[flink-dist-1.17-vvr-8.0.9-2-SNAPSHOT.
7. ...
8. Caused by: java.lang.RuntimeException: SplitFetcher thread 0 received unexpected exception while polling the records
9. at org.apache.flink.connector.base.source.reader.fetcher.SplitFetcher.runOnce(SplitFetcher.java:165) ~[flink-connector-files-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
10. at org.apache.flink.connector.base.source.reader.fetcher.SplitFetcher.run(SplitFetcher.java:114) ~[flink-connector-files-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
11. at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) ~[?:1.8.0_372]
12. at java.util.concurrent.FutureTask.run(FutureTask.java:266) ~[?:1.8.0_372]
13. at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) ~[?:1.8.0_372]
14. at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) ~[?:1.8.0_372]
15. ... 1 more
16. Caused by: java.lang.IllegalStateException: The connector is trying to read binlog starting at Struct{version=1.9.8.Final,connector=mysql,name=mysql_binlog_source,ts_ms=1732052840471,db=,server_id=0,file=mysql-bin.051880,pos=347695811,row=0}, but this is no longer available on the server. Reconfigure the connector to use a snapshot when needed.
17. at org.apache.flink.cdc.connectors.mysql.debezium.task.context.StatefulTaskContext.loadStartingOffsetState(StatefulTaskContext.java:212) ~[?:?]
18. at org.apache.flink.cdc.connectors.mysql.debezium.task.context.StatefulTaskContext.configure(StatefulTaskContext.java:133) ~[?:?]
19. at org.apache.flink.cdc.connectors.mysql.debezium.reader.BinlogSplitReader.submitSplit(BinlogSplitReader.java:105) ~[?:?]
20. at org.apache.flink.cdc.connectors.mysql.debezium.reader.BinlogSplitReader.submitSplit(BinlogSplitReader.java:71) ~[?:?]
21. at org.apache.flink.cdc.connectors.mysql.source.reader.MySqlSplitReader.pollSplitRecords(MySqlSplitReader.java:119) ~[?:?]
22. at org.apache.flink.cdc.connectors.mysql.source.reader.MySqlSplitReader.fetch(MySqlSplitReader.java:90) ~[?:?]
23. at org.apache.flink.connector.base.source.reader.fetcher.FetchTask.run(FetchTask.java:58) ~[flink-connector-files-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
24. at org.apache.flink.connector.base.source.reader.fetcher.SplitFetcher.runOnce(SplitFetcher.java:162) ~[flink-connector-files-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
25. at org.apache.flink.connector.base.source.reader.fetcher.SplitFetcher.run(SplitFetcher.java:114) ~[flink-connector-files-1.17-vvr-8.0.9-2-SNAPSHOT.jar:1.17-vvr-8.0.9-2-SNAPSHOT]
26. at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) ~[?:1.8.0_372]
27. at java.util.concurrent.FutureTask.run(FutureTask.java:266) ~[?:1.8.0_372]
28. at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) ~[?:1.8.0_372]
29. at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) ~[?:1.8.0_372]
30. ... 1 more
复制代码
办理：
参考debezium connectors和阿里云,增加如下参数：
1. debezium.max.queue.size: 162580
2. debezium.max.batch.size: 40960
3. debezium.poll.interval.ms: 50
4. scan.only.deserialize.captured.tables.changelog.enabled: true
复制代码
增量读取过慢，导致binlog 已经没了
参考阿里云,增加如下参数
1. scan.parallel-deserialize-changelog.enabled: true
2. scan.parallel-deserialize-changelog.handler.size: 4
3. heartbeat.interval: 5s
复制代码

免责声明：如果侵犯了您的权益，请联系站长，我们会及时删除侵权内容，谢谢合作！更多信息从访问主页：qidao123.com:ToB企服之家，中国第一个企服评测及商务社交产业平台。

欢迎光临 ToB企服应用市场:ToB评测及商务社交产业平台 (https://dis.qidao123.com/)