scala - 如何修复 "MetadataFetchFailedException: Missing an output location for shuffle"？

Question

如果我增加 word2vec 模型的模型大小，我开始在 log 中遇到这种异常:

org.apache.spark.shuffle.MetadataFetchFailedException: Missing an output location for shuffle 6
    at org.apache.spark.MapOutputTracker$$anonfun$org$apache$spark$MapOutputTracker$$convertMapStatuses$2.apply(MapOutputTracker.scala:542)
    at org.apache.spark.MapOutputTracker$$anonfun$org$apache$spark$MapOutputTracker$$convertMapStatuses$2.apply(MapOutputTracker.scala:538)
    at scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:772)
    at scala.collection.IndexedSeqOptimized$class.foreach(IndexedSeqOptimized.scala:33)
    at scala.collection.mutable.ArrayOps$ofRef.foreach(ArrayOps.scala:108)
    at scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:771)
    at org.apache.spark.MapOutputTracker$.org$apache$spark$MapOutputTracker$$convertMapStatuses(MapOutputTracker.scala:538)
    at org.apache.spark.MapOutputTracker.getMapSizesByExecutorId(MapOutputTracker.scala:155)
    at org.apache.spark.shuffle.BlockStoreShuffleReader.read(BlockStoreShuffleReader.scala:47)
    at org.apache.spark.rdd.ShuffledRDD.compute(ShuffledRDD.scala:98)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:270)
    at org.apache.spark.rdd.CoalescedRDD$$anonfun$compute$1.apply(CoalescedRDD.scala:96)
    at org.apache.spark.rdd.CoalescedRDD$$anonfun$compute$1.apply(CoalescedRDD.scala:95)
    at scala.collection.Iterator$$anon$13.hasNext(Iterator.scala:371)
    at scala.collection.Iterator$$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.ArrayBuffer.$plus$plus$eq(ArrayBuffer.scala:103)
    at scala.collection.mutable.ArrayBuffer.$plus$plus$eq(ArrayBuffer.scala:47)
    at scala.collection.TraversableOnce$class.to(TraversableOnce.scala:273)
    at scala.collection.AbstractIterator.to(Iterator.scala:1157)
    at scala.collection.TraversableOnce$class.toBuffer(TraversableOnce.scala:265)
    at scala.collection.AbstractIterator.toBuffer(Iterator.scala:1157)
    at scala.collection.TraversableOnce$class.toArray(TraversableOnce.scala:252)
    at scala.collection.AbstractIterator.toArray(Iterator.scala:1157)
    at org.apache.spark.rdd.RDD$$anonfun$collect$1$$anonfun$12.apply(RDD.scala:927)
    at org.apache.spark.rdd.RDD$$anonfun$collect$1$$anonfun$12.apply(RDD.scala:927)
    at org.apache.spark.SparkContext$$anonfun$runJob$5.apply(SparkContext.scala:1858)
    at org.apache.spark.SparkContext$$anonfun$runJob$5.apply(SparkContext.scala:1858)
    at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:66)
    at org.apache.spark.scheduler.Task.run(Task.scala:89)
    at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:213)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
    at java.lang.Thread.run(Thread.java:745)

我尝试编写自己的“保存模型”版本，如下所示:

  def save(model: Word2VecModel, sc: SparkContext, path: String): Unit = {

    println("Saving model as CSV ..")

    val vectorSize = model.getVectors.values.head.size

    println("vectorSize="+vectorSize)

    val SEPARATOR_TOKEN = " "    
    val dataArray = model.getVectors.toSeq.map { case (w, v) => Data(w, v) }

    println("Got dataArray ..")
    println("parallelize(dataArray, 10)")
    val par = sc.parallelize(dataArray, 10)
          .map(d => {

            val sb = new mutable.StringBuilder()
            sb.append(d.word)
            sb.append(SEPARATOR_TOKEN)

            for(v <- d.vector) {
              sb.append(v)
              sb.append(SEPARATOR_TOKEN)
            }
            sb.setLength(sb.length - 1)
            sb.append("\n")
            sb.toString()
          })
    println("repartition(1)")
    val rep = par.repartition(1)
    println("collect()")
    val vectorsAsString = rep.collect()

    println("Collected serialized vectors ..")    

    val cfile = new mutable.StringBuilder()

    cfile.append(vectorsAsString.length)
    cfile.append(" ")
    cfile.append(vectorSize)
    cfile.append("\n")

    val sb = new StringBuilder
    sb.append("word,")
    for(i <- 0 until vectorSize) {
      sb.append("v")
      sb.append(i.toString)
      sb.append(",")
    }
    sb.setLength(sb.length - 1)
    sb.append("\n")

    for(vectorString <- vectorsAsString) {
      sb.append(vectorString)
      cfile.append(vectorString)
    }

    println("Saving file to " + new Path(path, "data").toUri.toString)
    sc.parallelize(sb.toString().split("\n"), 1).saveAsTextFile(new Path(path+".csv", "data").toUri.toString)
    sc.parallelize(cfile.toString().split("\n"), 1).saveAsTextFile(new Path(path+".cs", "data").toUri.toString)
  }

显然它的工作原理类似于他们的 current implementation - 没有。

我想要一个 word2vec 模型。它适用于小文件，但不适用于模型变大的情况。

Answer 1

MetadataFetchFailedException当 MapOutputTracker 时抛出在执行程序上找不到本地缓存中分区的请求的随机映射输出，并尝试从驱动程序的 MapOutputTracker 远程获取它们.

这可能会导致几个结论:

驱动程序的内存问题

执行者的内存问题

执行者丢失

请查看日志以查找报告为“Executor lost”INFO 消息的问题和/或查看 Web UI 的 Executors 页面并查看 executors 是如何工作的。

executor 丢失的根本原因也可能是集群管理器决定杀死行为不端的 executor(可能使用了比请求更多的内存)。

看另一个问题 FetchFailedException or MetadataFetchFailedException when processing big data set了解更多信息。