This problem happen in Spark 1.3.1.  It happen when two jobs are running simultaneously each in its own Spark Context.

 

I don’t remember seeing this bug in Spark 1.2.0. Is it a new bug introduced in Spark 1.3.1?

 

Ningjun

 

From: Ted Yu [mailto:yuzhihong@gmail.com]
Sent: Wednesday, May 06, 2015 11:32 AM
To: Wang, Ningjun (LNG-NPV)
Cc: user@spark.apache.org
Subject: Re: java.io.IOException: org.apache.spark.SparkException: Failed to get broadcast_2_piece0

 

Which release of Spark are you using ?

 

Thanks


On May 6, 2015, at 8:03 AM, Wang, Ningjun (LNG-NPV) <ningjun.wang@lexisnexis.com> wrote:

I run a job on spark standalone cluster and got the exception below

 

Here is the line of code that cause problem

 

val myRdd: RDD[(String, String, String)] = … // RDD of (docid, cattegory, path)


myRdd.persist(StorageLevel.MEMORY_AND_DISK_SER)

val cats: Array[String] = myRdd.map(t => t._2).distinct().collect()  // This line cause the exception

 

 

15/05/06 10:48:51 WARN TaskSetManager: Lost task 0.0 in stage 0.0 (TID 0, LAB4-WIN03.pcc.lexisnexis.com): java.io.IOException: org.apache.spark.SparkException: Failed to get broadcast_2_piece0 of broadcast_2

        at org.apache.spark.util.Utils$.tryOrIOException(Utils.scala:1156)

        at org.apache.spark.broadcast.TorrentBroadcast.readBroadcastBlock(TorrentBroadcast.scala:164)

        at org.apache.spark.broadcast.TorrentBroadcast._value$lzycompute(TorrentBroadcast.scala:64)

        at org.apache.spark.broadcast.TorrentBroadcast._value(TorrentBroadcast.scala:64)

        at org.apache.spark.broadcast.TorrentBroadcast.getValue(TorrentBroadcast.scala:87)

        at org.apache.spark.broadcast.Broadcast.value(Broadcast.scala:70)

        at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:61)

        at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)

        at org.apache.spark.scheduler.Task.run(Task.scala:64)

        at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:203)

        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:745)

Caused by: org.apache.spark.SparkException: Failed to get broadcast_2_piece0 of broadcast_2

        at org.apache.spark.broadcast.TorrentBroadcast$$anonfun$org$apache$spark$broadcast$TorrentBr

oadcast$$readBlocks$1$$anonfun$2.apply(TorrentBroadcast.scala:137)

        at org.apache.spark.broadcast.TorrentBroadcast$$anonfun$org$apache$spark$broadcast$TorrentBr

oadcast$$readBlocks$1$$anonfun$2.apply(TorrentBroadcast.scala:137)

        at scala.Option.getOrElse(Option.scala:120)

        at org.apache.spark.broadcast.TorrentBroadcast$$anonfun$org$apache$spark$broadcast$TorrentBr

oadcast$$readBlocks$1.apply$mcVI$sp(TorrentBroadcast.scala:136)

        at org.apache.spark.broadcast.TorrentBroadcast$$anonfun$org$apache$spark$broadcast$TorrentBr

oadcast$$readBlocks$1.apply(TorrentBroadcast.scala:119)

        at org.apache.spark.broadcast.TorrentBroadcast$$anonfun$org$apache$spark$broadcast$TorrentBr

oadcast$$readBlocks$1.apply(TorrentBroadcast.scala:119)

        at scala.collection.immutable.List.foreach(List.scala:318)

        at org.apache.spark.broadcast.TorrentBroadcast.org$apache$spark$broadcast$TorrentBroadcast$$

readBlocks(TorrentBroadcast.scala:119)

        at org.apache.spark.broadcast.TorrentBroadcast$$anonfun$readBroadcastBlock$1.apply(TorrentBr

oadcast.scala:174)

        at org.apache.spark.util.Utils$.tryOrIOException(Utils.scala:1153)

        ... 12 more

 

 

Any idea what cause the problem and how to avoid it?

 

Thanks

Ningjun