[SPARK-28009] PipedRDD: Block not locked for reading failure

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 2.4.0
• Component/s: Spark Core

Environment

Running in a Docker container with Spark 2.4.0 on Linux kernel 4.9.0

• Assignee: Unassigned
• Reporter: Douglas Colkitt
• Created: 11/Jun/19 23:47
• Updated: 13/Jun/19 03:24
• Resolved: 13/Jun/19 03:24

Description

PipedRDD operation fails with the below stack trace. Failure primarily occurs when the STDOUT from the Unix process is small and the STDIN into the Unix process is comparatively much larger.

Given the similarity to SPARK-18406, this seems to be due to a race condition when it comes to accessing the block's reader locker. The PipedRDD class implementation spawns STDIN iterator in a separate thread, so that would corroborate the race condition hypothesis.

at scala.Predef$.assert(Predef.scala:170)
at org.apache.spark.storage.BlockInfoManager.unlock(BlockInfoManager.scala:299)
at org.apache.spark.storage.BlockManager.releaseLock(BlockManager.scala:842)
at org.apache.spark.storage.BlockManager.releaseLockAndDispose(BlockManager.scala:1610)
at org.apache.spark.storage.BlockManager$$anonfun$2.apply$mcV$sp(BlockManager.scala:621)
at org.apache.spark.util.CompletionIterator$$anon$1.completion(CompletionIterator.scala:44)
at org.apache.spark.util.CompletionIterator.hasNext(CompletionIterator.scala:33)
at org.apache.spark.InterruptibleIterator.hasNext(InterruptibleIterator.scala:37)
at scala.collection.Iterator$$anon$13.hasNext(Iterator.scala:462)
at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:409)
at scala.collection.Iterator$class.foreach(Iterator.scala:891)
at scala.collection.AbstractIterator.foreach(Iterator.scala:1334)
at org.apache.spark.rdd.PipedRDD$$anon$3.run(PipedRDD.scala:145)
Suppressed: java.lang.AssertionError: assertion failed
at scala.Predef$.assert(Predef.scala:156)
at org.apache.spark.storage.BlockInfo.checkInvariants(BlockInfoManager.scala:84)
at org.apache.spark.storage.BlockInfo.readerCount_$eq(BlockInfoManager.scala:66)
at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:363)
at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:362)
at scala.Option.foreach(Option.scala:257)
at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:362)
at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:358)
at scala.collection.Iterator$class.foreach(Iterator.scala:891)
at scala.collection.AbstractIterator.foreach(Iterator.scala:1334)
at org.apache.spark.storage.BlockInfoManager.releaseAllLocksForTask(BlockInfoManager.scala:358)
at org.apache.spark.storage.BlockManager.releaseAllLocksForTask(BlockManager.scala:858)
at org.apache.spark.executor.Executor$TaskRunner$$anonfun$1.apply$mcV$sp(Executor.scala:409)
at org.apache.spark.util.Utils$.tryWithSafeFinally(Utils.scala:1369)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:408)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)

Issue Links

• duplicates: SPARK-27666 Do not release lock while TaskContext already completed