Details
-
Bug
-
Status: Resolved
-
Blocker
-
Resolution: Duplicate
-
1.2.0
-
None
-
None
Description
OOM/GC errors with sort-based shuffle that go away when shuffle.manager is set to hash. Also observing memory leak issues. Can be reproduced with PageRank with a large enough input dataset. Sample stack trace:
19:22:14 WARN TaskSetManager: Lost task 56.0 in stage 1.0 (TID 298, c1706.halxg.cloudera.com): java.lang.OutOfMemoryError: Java heap space com.ning.compress.BufferRecycler.allocDecodeBuffer(BufferRecycler.java:137) com.ning.compress.lzf.LZFInputStream.<init>(LZFInputStream.java:111) com.ning.compress.lzf.LZFInputStream.<init>(LZFInputStream.java:97) com.ning.compress.lzf.LZFInputStream.<init>(LZFInputStream.java:80) org.apache.spark.io.LZFCompressionCodec.compressedInputStream(CompressionCodec.scala:107) org.apache.spark.storage.BlockManager.wrapForCompression(BlockManager.scala:1155) org.apache.spark.util.collection.ExternalSorter$SpillReader.nextBatchStream(ExternalSorter.scala:559) org.apache.spark.util.collection.ExternalSorter$SpillReader.<init>(ExternalSorter.scala:531) org.apache.spark.util.collection.ExternalSorter$$anonfun$6.apply(ExternalSorter.scala:379) org.apache.spark.util.collection.ExternalSorter$$anonfun$6.apply(ExternalSorter.scala:379) scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244) scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59) scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:47) scala.collection.TraversableLike$class.map(TraversableLike.scala:244) scala.collection.AbstractTraversable.map(Traversable.scala:105) org.apache.spark.util.collection.ExternalSorter.merge(ExternalSorter.scala:379) org.apache.spark.util.collection.ExternalSorter.partitionedIterator(ExternalSorter.scala:686) org.apache.spark.util.collection.ExternalSorter.writePartitionedFile(ExternalSorter.scala:739) org.apache.spark.shuffle.sort.SortShuffleWriter.write(SortShuffleWriter.scala:70) org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:68) org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41) org.apache.spark.scheduler.Task.run(Task.scala:56) org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:195) java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145) java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615) java.lang.Thread.run(Thread.java:745)
Attachments
Issue Links
- duplicates
-
SPARK-4467 Number of elements read is never reset in ExternalSorter
- Closed