scala - Spark 在大型洗牌作业上失败，出现 java.io.IOException : Filesystem closed

Question

我经常发现 spark 在处理大型作业时失败，并出现无用的无意义异常。工作日志看起来正常，没有错误，但它们的状态为“KILLED”。这对于大型随机播放非常常见，因此像 .distinct 这样的操作。

问题是，我如何诊断出了什么问题，理想情况下，我该如何修复它？

考虑到这些操作中有很多是幺半群的，我一直在通过将数据分成例如 10 个 block ，在每个 block 上运行应用程序，然后在所有结果输出上运行应用程序来解决这个问题。换句话说 - 元映射减少。

14/06/04 12:56:09 ERROR client.AppClient$ClientActor: Master removed our application: FAILED; stopping client
14/06/04 12:56:09 WARN cluster.SparkDeploySchedulerBackend: Disconnected from Spark cluster! Waiting for reconnection...
14/06/04 12:56:09 WARN scheduler.TaskSetManager: Loss was due to java.io.IOException
java.io.IOException: Filesystem closed
    at org.apache.hadoop.hdfs.DFSClient.checkOpen(DFSClient.java:703)
    at org.apache.hadoop.hdfs.DFSInputStream.readWithStrategy(DFSInputStream.java:779)
    at org.apache.hadoop.hdfs.DFSInputStream.read(DFSInputStream.java:840)
    at java.io.DataInputStream.read(DataInputStream.java:149)
    at org.apache.hadoop.io.compress.DecompressorStream.getCompressedData(DecompressorStream.java:159)
    at org.apache.hadoop.io.compress.DecompressorStream.decompress(DecompressorStream.java:143)
    at org.apache.hadoop.io.compress.DecompressorStream.read(DecompressorStream.java:85)
    at java.io.InputStream.read(InputStream.java:101)
    at org.apache.hadoop.util.LineReader.fillBuffer(LineReader.java:180)
    at org.apache.hadoop.util.LineReader.readDefaultLine(LineReader.java:216)
    at org.apache.hadoop.util.LineReader.readLine(LineReader.java:174)
    at org.apache.hadoop.mapred.LineRecordReader.next(LineRecordReader.java:209)
    at org.apache.hadoop.mapred.LineRecordReader.next(LineRecordReader.java:47)
    at org.apache.spark.rdd.HadoopRDD$$anon$1.getNext(HadoopRDD.scala:164)
    at org.apache.spark.rdd.HadoopRDD$$anon$1.getNext(HadoopRDD.scala:149)
    at org.apache.spark.util.NextIterator.hasNext(NextIterator.scala:71)
    at org.apache.spark.InterruptibleIterator.hasNext(InterruptibleIterator.scala:27)
    at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
    at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
    at scala.collection.Iterator$class.foreach(Iterator.scala:727)
    at scala.collection.AbstractIterator.foreach(Iterator.scala:1157)
    at scala.collection.generic.Growable$class.$plus$plus$eq(Growable.scala:48)
    at scala.collection.mutable.ListBuffer.$plus$plus$eq(ListBuffer.scala:176)
    at scala.collection.mutable.ListBuffer.$plus$plus$eq(ListBuffer.scala:45)
    at scala.collection.TraversableOnce$class.to(TraversableOnce.scala:273)
    at scala.collection.AbstractIterator.to(Iterator.scala:1157)
    at scala.collection.TraversableOnce$class.toList(TraversableOnce.scala:257)
    at scala.collection.AbstractIterator.toList(Iterator.scala:1157)
    at $line5.$read$$iwC$$iwC$$iwC$$iwC$$anonfun$2.apply(<console>:13)
    at $line5.$read$$iwC$$iwC$$iwC$$iwC$$anonfun$2.apply(<console>:13)
    at org.apache.spark.rdd.RDD$$anonfun$1.apply(RDD.scala:450)
    at org.apache.spark.rdd.RDD$$anonfun$1.apply(RDD.scala:450)
    at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:34)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:241)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:232)
    at org.apache.spark.rdd.MappedRDD.compute(MappedRDD.scala:31)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:241)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:232)
    at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:34)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:241)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:232)
    at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:161)
    at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:102)
    at org.apache.spark.scheduler.Task.run(Task.scala:53)
    at org.apache.spark.executor.Executor$TaskRunner$$anonfun$run$1.apply$mcV$sp(Executor.scala:213)
    at org.apache.spark.deploy.SparkHadoopUtil$$anon$1.run(SparkHadoopUtil.scala:42)
    at org.apache.spark.deploy.SparkHadoopUtil$$anon$1.run(SparkHadoopUtil.scala:41)
    at java.security.AccessController.doPrivileged(Native Method)
    at javax.security.auth.Subject.doAs(Subject.java:415)
    at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1548)
    at org.apache.spark.deploy.SparkHadoopUtil.runAsUser(SparkHadoopUtil.scala:41)
    at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:178)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
    at java.lang.Thread.run(Thread.java:744)

Answer 1

截至 2014 年 9 月 1 日，这是 Spark 中的“开放式改进”。请看https://issues.apache.org/jira/browse/SPARK-3052 .正如 syrza 在给定链接中指出的那样，当执行程序失败导致此消息时，关闭 Hook 可能以错误的顺序完成。我知道您将需要进行更多调查才能找出问题的主要原因(即执行人失败的原因)。如果它是一个大的洗牌，它可能是一个内存不足的错误，它会导致执行程序失败，然后导致 Hadoop 文件系统在它们的关闭 Hook 中关闭。因此，该执行程序的运行任务中的 RecordReader 抛出“java.io.IOException: Filesystem closed”异常。我猜它会在后续版本中修复，然后您会收到更多有用的错误消息:)