[
https://issues.apache.org/jira/browse/SPARK-12089?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15036327#comment-15036327
]
Erik Selin commented on SPARK-12089:
------------------------------------
It's a bunch of table joins followed by a group by on multiple fields. One of
the fields being a month window. Example:
{code}
select x.a, x.b, x.c, x.month from (
select a, b, c, concat(year(from_unixtime(t)), "-", month(from_unixtime(t)))
AS month
from foo
left join bar on foo.bar_id = bar.id
left join biz on foo.biz_id = biz.id
) as x
group by x.a, x.b, x.c, x.month
{code}
My hypothesis, but I need your expertise to confirm/deny since I'm really not
familiar with this area of sparks codebase. Is that the monthly group by is
indeed creating something huge, perhaps by bucketing a lot of data together
into the same buffer? I have similar jobs running very similar queries but
grouped by day and they are not running into this issue.
I'll do a debug log run once I have some spare cycles on my end! :)
> java.lang.NegativeArraySizeException when growing BufferHolder
> --------------------------------------------------------------
>
> Key: SPARK-12089
> URL: https://issues.apache.org/jira/browse/SPARK-12089
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 1.6.0
> Reporter: Erik Selin
> Priority: Critical
>
> When running a large spark sql query including multiple joins I see tasks
> failing with the following trace:
> {code}
> java.lang.NegativeArraySizeException
> at
> org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder.grow(BufferHolder.java:36)
> at
> org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter.write(UnsafeRowWriter.java:188)
> at
> org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificUnsafeProjection.apply(Unknown
> Source)
> at
> org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificUnsafeProjection.apply(Unknown
> Source)
> at
> org.apache.spark.sql.execution.joins.OneSideOuterIterator.getRow(SortMergeOuterJoin.scala:288)
> at
> org.apache.spark.sql.execution.RowIteratorToScala.next(RowIterator.scala:76)
> at
> org.apache.spark.sql.execution.RowIteratorToScala.next(RowIterator.scala:62)
> at scala.collection.Iterator$$anon$11.next(Iterator.scala:328)
> at
> org.apache.spark.shuffle.sort.UnsafeShuffleWriter.write(UnsafeShuffleWriter.java:164)
> at
> org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:73)
> at
> org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)
> at org.apache.spark.scheduler.Task.run(Task.scala:88)
> at
> org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:214)
> 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)
> {code}
> From the spark code it looks like this is due to a integer overflow when
> growing a buffer length. The offending line {{BufferHolder.java:36}} is the
> following in the version I'm running:
> {code}
> final byte[] tmp = new byte[length * 2];
> {code}
> This seems to indicate to me that this buffer will never be able to hold more
> then 2G worth of data. And likely will hold even less since any length >
> 1073741824 will cause a integer overflow and turn the new buffer size
> negative.
> I hope I'm simply missing some critical config setting but it still seems
> weird that we have a (rather low) upper limit on these buffers.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
