spark-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From <>
Subject RE: getting Caused by: org.apache.spark.SparkException: Job failed: Task 1.0:1 failed more than 4 times
Date Wed, 23 Oct 2013 22:51:32 GMT
No, the problem was worked around by not caching many RDD's in my program.  For example not
caching col_values.distinct() then collect works as expected.

I know this cause I was tracking the invocation of each call with system out and so if FileUtil.extractValueForAPositionNo
is failing it will get displayed in the output.

From: Mark Hamstra []
Sent: Wednesday, October 23, 2013 3:00 PM
To: user
Subject: Re: getting Caused by: org.apache.spark.SparkException: Job failed: Task 1.0:1 failed
more than 4 times

When you say things like "apply map works" and then lay the blame for the job failure on collect(),
that's not being fair to collect().  RDD transformations are lazy, so the code that you posted
immediately after claiming that map works doesn't really do anything right then and there
except to schedule the map transformation to be done at a later time when an RDD action is
invoked -- collect() in this case.  Very likely FileUtil.extractValueForAPositionNo(line,
columnPosition) is throwing an exception, which causes the job initiated by the collect()
action to fail.

On Wed, Oct 23, 2013 at 2:02 PM, <<>>
I have spark 0.8.0 running in cluster with 2 workers each setup with 16 cores and 24GB memory
against hadoop 1.2.1

I have csv with over 1 million records.

My spark jave program runs as expected with smaller size csv but fails as follows:

Loading csv as text works
              JavaRDD<String> rawTable = sc.textFile(raw_file_path).cache();

Then apply map works
              JavaRDD<String> col_values =
                           new Function<String, String>() {
                                  private static final long serialVersionUID = 1L;

                                  public String call(String line) throws Exception {
                                         return FileUtil.extractValueForAPositionNo(line,

Then getting distinct works also
                                  JavaRDD<String> distinct_col_values = col_values.distinct().cache();

But to dump the content of the distinct into an List of String object .. fails
                                  List<String> list = distinct_col_values.collect();

Any help?

        at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
        at sun.reflect.NativeMethodAccessorImpl.invoke(
        at sun.reflect.DelegatingMethodAccessorImpl.invoke(
        at java.lang.reflect.Method.invoke(
        at org.codehaus.mojo.exec.ExecJavaMojo$
Caused by: org.apache.spark.SparkException: Job failed: Task 1.0:1 failed more than 4 times
        at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:760)
        at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:758)
        at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:60)
        at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:47)
        at org.apache.spark.scheduler.DAGScheduler.abortStage(DAGScheduler.scala:758)
        at org.apache.spark.scheduler.DAGScheduler.processEvent(DAGScheduler.scala:379)
        at org.apache.spark.scheduler.DAGScheduler$$anon$

View raw message