我正试图在Azure HDInsight按需集群上运行一个基于Spark的应用程序,并看到许多SparkException(由ConcurrentModificationException引起)被记录。启动本地Spark实例时,应用程序运行时没有这些错误。
我看过关于
similar errors when using accumulators
我的代码确实使用了CollectionAccumulator,但是我在使用它的任何地方都放置了同步块,这没有什么区别。与累加器相关的代码如下所示:
class MySparkClass(sc : SparkContext) {
val myAccumulator = sc.collectionAccumulator[MyRecord]
override def add(record: MyRecord) = {
synchronized {
myAccumulator.add(record)
}
}
override def endOfBatch() = {
synchronized {
myAccumulator.value.asScala.foreach((record: MyRecord) => {
processIt(record)
})
}
}
}
但是,异常不会导致应用程序失败
endOfBatch
调用时,代码尝试从累加器中读取值。累加器为空且
processIt
从来没有人打过电话。
我们正在使用
HDInsight version 3.6
使用Spark 2.3.0版
18/11/26 11:04:37 WARN Executor: Issue communicating with driver in heartbeater
org.apache.spark.SparkException: Exception thrown in awaitResult:
at org.apache.spark.util.ThreadUtils$.awaitResult(ThreadUtils.scala:205)
at org.apache.spark.rpc.RpcTimeout.awaitResult(RpcTimeout.scala:75)
at org.apache.spark.rpc.RpcEndpointRef.askSync(RpcEndpointRef.scala:92)
at org.apache.spark.executor.Executor.org$apache$spark$executor$Executor$$reportHeartBeat(Executor.scala:785)
at org.apache.spark.executor.Executor$$anon$2$$anonfun$run$1.apply$mcV$sp(Executor.scala:814)
at org.apache.spark.executor.Executor$$anon$2$$anonfun$run$1.apply(Executor.scala:814)
at org.apache.spark.executor.Executor$$anon$2$$anonfun$run$1.apply(Executor.scala:814)
at org.apache.spark.util.Utils$.logUncaughtExceptions(Utils.scala:1988)
at org.apache.spark.executor.Executor$$anon$2.run(Executor.scala:814)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: java.util.ConcurrentModificationException
at java.util.ArrayList.writeObject(ArrayList.java:770)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at java.io.ObjectStreamClass.invokeWriteObject(ObjectStreamClass.java:1140)
at java.io.ObjectOutputStream.writeSerialData(ObjectOutputStream.java:1496)
at java.io.ObjectOutputStream.writeOrdinaryObject(ObjectOutputStream.java:1432)
at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1178)
at java.io.ObjectOutputStream.defaultWriteFields(ObjectOutputStream.java:1548)
at java.io.ObjectOutputStream.defaultWriteObject(ObjectOutputStream.java:441)
at java.util.Collections$SynchronizedCollection.writeObject(Collections.java:2081)
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 java.io.ObjectStreamClass.invokeWriteObject(ObjectStreamClass.java:1140)
at java.io.ObjectOutputStream.writeSerialData(ObjectOutputStream.java:1496)
at java.io.ObjectOutputStream.writeOrdinaryObject(ObjectOutputStream.java:1432)
at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1178)
at java.io.ObjectOutputStream.defaultWriteFields(ObjectOutputStream.java:1548)
at java.io.ObjectOutputStream.writeSerialData(ObjectOutputStream.java:1509)
at java.io.ObjectOutputStream.writeOrdinaryObject(ObjectOutputStream.java:1432)
at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1178)
at java.io.ObjectOutputStream.writeArray(ObjectOutputStream.java:1378)
at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1174)
at java.io.ObjectOutputStream.defaultWriteFields(ObjectOutputStream.java:1548)
at java.io.ObjectOutputStream.writeSerialData(ObjectOutputStream.java:1509)
at java.io.ObjectOutputStream.writeOrdinaryObject(ObjectOutputStream.java:1432)
at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1178)
at java.io.ObjectOutputStream.defaultWriteFields(ObjectOutputStream.java:1548)
at java.io.ObjectOutputStream.writeSerialData(ObjectOutputStream.java:1509)
at java.io.ObjectOutputStream.writeOrdinaryObject(ObjectOutputStream.java:1432)
at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1178)
at java.io.ObjectOutputStream.writeArray(ObjectOutputStream.java:1378)
at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1174)
at java.io.ObjectOutputStream.defaultWriteFields(ObjectOutputStream.java:1548)
at java.io.ObjectOutputStream.writeSerialData(ObjectOutputStream.java:1509)
at java.io.ObjectOutputStream.writeOrdinaryObject(ObjectOutputStream.java:1432)
at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1178)
at java.io.ObjectOutputStream.writeObject(ObjectOutputStream.java:348)
at org.apache.spark.serializer.JavaSerializationStream.writeObject(JavaSerializer.scala:43)
at org.apache.spark.rpc.netty.RequestMessage.serialize(NettyRpcEnv.scala:565)
at org.apache.spark.rpc.netty.NettyRpcEnv.ask(NettyRpcEnv.scala:231)
at org.apache.spark.rpc.netty.NettyRpcEndpointRef.ask(NettyRpcEnv.scala:523)
at org.apache.spark.rpc.RpcEndpointRef.askSync(RpcEndpointRef.scala:91)
... 13 more
下面的代码是一个更加独立的示例,重现了这个问题。
MyRecord
是一个只包含数值的简单case类。代码在本地运行时不会出错,但在HDInsight集群上会产生上述错误。
object MainDemo {
def main(args: Array[String]) {
val sparkContext = SparkSession.builder.master("local[4]").getOrCreate().sparkContext
val myAccumulator = sparkContext.collectionAccumulator[MyRecord]
sparkContext.binaryFiles("/my/files/here").foreach(_ => {
for(i <- 1 to 100000) {
val record = MyRecord(i, 0, 0)
myAccumulator.add(record)
}
})
myAccumulator.value.asScala.foreach((record: MyRecord) => {
// we expect this to be called once for each record that we 'add' above,
// but it is never called
println(record)
})
}
}
