背景
我的光束管道被设计为处理Avro SpecificRecordBase类型的元素。
为了简化我的问题,假设我有两种以Avro格式生成的元素,它们都有各自的字段:
class Dog extends SpecificRecordBase {
....
}
class Cat extends SpecificRecordBase {
...
}
管道将从输入的Kafka中读取元素,处理元素并将已处理的元素放入输出Kafka中,如下所示:
Pipeline pipeline = Pipeline.create(getOptions());
pipeline.getCoderRegistry().registerCoderForClass(SpecificRecordBase.class, <what shall I put here?>);
pipeline.apply(kafkaReaderTransformer)
.apply(Window.into(FixedWindows.of(Duration.standardSeconds(getWindowSize()))))
.apply(GroupByKey.create())
.apply(ParDo.of(GiveShowerToPetDoFn))
.apply(Flatten.iterables())
.apply(kafkaWriterTransformer);
问题
我的问题是如何在管道中注册编码器?由于将来可以从Cat Kafka或Dog Kafka甚至Toad Kafka中读取pipleline,因此我需要一种通用的方式来注册编码器,该编码器可以序列化在运行时确定的SpecificRecordBase的所有子类。
我的失败解决方案
我尝试了以下代码来填充代码中的空白:
AvroCoder.of(SpecificRecordBase.class):不起作用
我在运行管道时出现以下错误:
Caused by: avro.shaded.com.google.common.util.concurrent.UncheckedExecutionException: org.apache.avro.AvroRuntimeException: Not a Specific class: class org.apache.avro.specific.SpecificRecordBase
at avro.shaded.com.google.common.cache.LocalCache$Segment.get(LocalCache.java:2234)
at avro.shaded.com.google.common.cache.LocalCache.get(LocalCache.java:3965)
at avro.shaded.com.google.common.cache.LocalCache.getOrLoad(LocalCache.java:3969)
at avro.shaded.com.google.common.cache.LocalCache$LocalManualCache.get(LocalCache.java:4829)
at org.apache.avro.specific.SpecificData.getSchema(SpecificData.java:225)
... 23 more
Caused by: org.apache.avro.AvroRuntimeException: Not a Specific class: class org.apache.avro.specific.SpecificRecordBase
at org.apache.avro.specific.SpecificData.createSchema(SpecificData.java:285)
at org.apache.avro.reflect.ReflectData.createSchema(ReflectData.java:594)
at org.apache.avro.specific.SpecificData$2.load(SpecificData.java:218)
at org.apache.avro.specific.SpecificData$2.load(SpecificData.java:215)
at avro.shaded.com.google.common.cache.LocalCache$LoadingValueReference.loadFuture(LocalCache.java:3568)
at avro.shaded.com.google.common.cache.LocalCache$Segment.loadSync(LocalCache.java:2350)
at avro.shaded.com.google.common.cache.LocalCache$Segment.lockedGetOrLoad(LocalCache.java:2313)
at avro.shaded.com.google.common.cache.LocalCache$Segment.get(LocalCache.java:2228)
... 27 more
SerializableCoder.of(SpecificRecordBase.class):引发令人困惑的异常
这应该是一个有前途的选择,但是当我运行Pipeline时,我在下面遇到了一个非常令人困惑的错误,以下内容令人困惑,因为Cat实际上是通过继承自SpecificRecordBase实现可序列化的:
Caused by: java.lang.ClassCastException: Cat cannot be cast to java.io.Serializable
at org.apache.beam.sdk.coders.SerializableCoder.encode(SerializableCoder.java:53)
at org.apache.beam.sdk.coders.Coder.encode(Coder.java:136)
at org.apache.beam.sdk.util.WindowedValue$FullWindowedValueCoder.encode(WindowedValue.java:578)
at org.apache.beam.sdk.util.WindowedValue$FullWindowedValueCoder.encode(WindowedValue.java:569)
at org.apache.beam.sdk.util.WindowedValue$FullWindowedValueCoder.encode(WindowedValue.java:529)
at org.apache.beam.runners.spark.coders.CoderHelpers.toByteArray(CoderHelpers.java:53)
at org.apache.beam.runners.spark.coders.CoderHelpers.lambda$toByteFunction$28e77fe8$1(CoderHelpers.java:143)
at org.apache.spark.api.java.JavaPairRDD$$anonfun$pairFunToScalaFun$1.apply(JavaPairRDD.scala:1043)
at org.apache.spark.api.java.JavaPairRDD$$anonfun$pairFunToScalaFun$1.apply(JavaPairRDD.scala:1043)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:409)
at org.apache.spark.shuffle.sort.BypassMergeSortShuffleWriter.write(BypassMergeSortShuffleWriter.java:149)
at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:96)
at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:53)
at org.apache.spark.scheduler.Task.run(Task.scala:109)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:345)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.IllegalStateException: Unable to return a default Coder for ParDo(Deserialize)/ParMultiDo(Deserialize).output [PCollection]. Correct one of the following root causes:
No Coder has been manually specified; you may do so using .setCoder().
Inferring a Coder from the CoderRegistry failed: Cannot provide coder for parameterized type org.apache.beam.sdk.values.KV<java.lang.String, org.apache.avro.specific.SpecificRecordBase>: Unable to provide a Coder for org.apache.avro.specific.SpecificRecordBase.
Building a Coder using a registered CoderProvider failed.
See suppressed exceptions for detailed failures.
Using the default output Coder from the producing PTransform failed: PTransform.getOutputCoder called.
at org.apache.beam.vendor.guava.v26_0_jre.com.google.common.base.Preconditions.checkState(Preconditions.java:507)
at org.apache.beam.sdk.values.PCollection.getCoder(PCollection.java:278)
at org.apache.beam.sdk.values.PCollection.finishSpecifying(PCollection.java:115)
at org.apache.beam.sdk.runners.TransformHierarchy.finishSpecifyingInput(TransformHierarchy.java:191)
at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:538)
at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:473)
at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:357)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at org.springframework.beans.factory.support.SimpleInstantiationStrategy.instantiate(SimpleInstantiationStrategy.java:154)
我终于使用解决方法解决了此处发布的问题。
根本原因
原来是由不同环境中的编码器不兼容引起的。尽管编码器在我的本地环境中工作,但是prod依赖版本有所不同,导致Beam库无法对从SpecificRecordBase派生的类进行编码和解码。
两种解决方案
1)使用字节作为输入并使用字节作为输出来更改管道中的每个doFun:
public class GiveShowerToPetDoFn extends DoFn<KV<String, byte[]>, KV<String, byte[]>> {
...
}
这意味着您将在执行实际的业务逻辑之前手动从字节反序列化对象,并在最后一步将结果序列化回字节。这使Beam使用adoli的doFun之间的默认字节编码器/解码器,并且该编码器/解码器将始终有效,因为它正在处理基本类型而不是自定义类型。
2)为您的自定义类型编写自己的编码器/解码器。
解决方案1和2本质上是相同的。就我而言,我使用第一个解决方案来解决我的问题。