spark-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Shangyu Luo <lsy...@gmail.com>
Subject Re: Questions about the files that Spark will produce during its running
Date Tue, 29 Oct 2013 06:28:43 GMT
Yes,
I broadcast the spark-env.sh file to all worker nodes before I run my
program and then execute bin/stop-all.sh, bin/start-all.sh.
I have also viewed the size of data2 directory on each worker node and it
is also about 800G.
Thanks!


2013/10/29 Matei Zaharia <matei.zaharia@gmail.com>

> The error is from a worker node -- did you check that /data2 is set up
> properly on the worker nodes too? In general that should be the only
> directory used.
>
> Matei
>
> On Oct 28, 2013, at 6:52 PM, Shangyu Luo <lsyurd@gmail.com> wrote:
>
> Hello,
> I have some questions about the files that Spark will create and use
> during its running.
> (1) I am running a python program on Spark with a cluster of EC2. The data
> comes from hdfs file system.  I have met the following error in the console
> of the master node:
> *java.io.FileNotFoundException:
> /data2/tmp/spark-local-20131029003412-c340/1b/shuffle_1_527_79 (No space
> left on device)*
>         at java.io.FileOutputStream.openAppend(Native Method)
>         at java.io.FileOutputStream.<init>(FileOutputStream.java:207)
>         at
> org.apache.spark.storage.DiskStore$DiskBlockObjectWriter.open(DiskStore.scala:58)
>         at
> org.apache.spark.storage.DiskStore$DiskBlockObjectWriter.write(DiskStore.scala:107)
>         at
> org.apache.spark.scheduler.ShuffleMapTask$$anonfun$run$1.apply(ShuffleMapTask.scala:152)
>         at
> org.apache.spark.scheduler.ShuffleMapTask$$anonfun$run$1.apply(ShuffleMapTask.scala:149)
>         at scala.collection.Iterator$class.foreach(Iterator.scala:772)
>         at scala.collection.Iterator$$anon$19.foreach(Iterator.scala:399)
>         at
> org.apache.spark.scheduler.ShuffleMapTask.run(ShuffleMapTask.scala:149)
>         at
> org.apache.spark.scheduler.ShuffleMapTask.run(ShuffleMapTask.scala:88)
>         at
> org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:158)
>         at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1146)
>         at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
>         at java.lang.Thread.run(Thread.java:679)
> I set spark.local.dir=*/data2/tmp* in spark-env.sh and there is about *
> 800G* space in data2 directory. I have checked the space of data2 and it
> is just used about 3G.
> So why Spark thinks that there is no space left on device?
>
> (2) Moreover, *I am wondering if Spark will create some files  under
> other directories other than spark.local.dir*? Presently I use a=b.map.*
> persist(storage.disk_only)* in some part of my program, where will the
> persisted data be stored?
>
> (3) Lastly, I also had "*Removing BlockManager xxx with no recent heart
> beats: xxxxxms exceeds 45000ms* " error sometimes. I have set the
> corresponding parameters in spark-env.sh:
> SPARK_JAVA_OPTS+="-Dspark.akka.timeout=300000 "
> SPARK_JAVA_OPTS+="-Dspark.worker.timeout=300000 "
> SPARK_JAVA_OPTS+="-Dspark.akka.askTimeout=3000 "
> SPARK_JAVA_OPTS+="-Dspark.storage.blockManagerHeartBeatMs=300000 "
> SPARK_JAVA_OPTS+="-Dspark.akka.retry.wait=300000 "
> But it does no help. Can someone gives me some suggestion about solving
> this problem?
>
> Any help will be appreciated!
> Thanks!
>
> Best,
> Shangyu
>
>
>


-- 
--

Shangyu, Luo
Department of Computer Science
Rice University

--
Not Just Think About It, But Do It!
--
Success is never final.
--
Losers always whine about their best

Mime
View raw message