spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Liwei Lin (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (SPARK-17842) Thread and memory leak in WindowDstream (UnionRDD ) when parallelPartition computation gets enabled.
Date Mon, 10 Oct 2016 08:42:20 GMT

    [ https://issues.apache.org/jira/browse/SPARK-17842?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15561099#comment-15561099
] 

Liwei Lin edited comment on SPARK-17842 at 10/10/16 8:41 AM:
-------------------------------------------------------------

-[~sreelalsl] thanks for the very clear reproducer -- it can be easily reproduced against
2.0 and master(as of 8a6bbe095b6a9aa33989c0deaa5ed0128d70320f).-

-I'll submit a patch.-

Update:

[~sreelalsl], this seems to be a duplicate to SPARK-17396, whose fix was rolled out in 2.0.1
just days ago. Can you try 2.0.1 to see if that works for you?


was (Author: lwlin):
[~sreelalsl] thanks for the very clear reproducer -- it can be easily reproduced against 2.0
and master(as of 8a6bbe095b6a9aa33989c0deaa5ed0128d70320f).

I'll submit a patch.

> Thread and memory leak in WindowDstream (UnionRDD ) when parallelPartition computation
gets enabled. 
> -----------------------------------------------------------------------------------------------------
>
>                 Key: SPARK-17842
>                 URL: https://issues.apache.org/jira/browse/SPARK-17842
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Core, Streaming
>    Affects Versions: 2.0.0
>         Environment: Yarn cluster, Eclipse Dev Env
>            Reporter: Sreelal S L
>            Priority: Critical
>
> We noticed a steady increase in ForkJoinTask instances in the driver process heap. Found
out the root cause to be UnionRDD.
> WindowDstream internally uses UnionRDD which has a parallel partition computation logic
by using parallel collection with ForkJoinPool task support. 
> partitionEvalTaskSupport =new ForkJoinTaskSupport(new ForkJoinPool(8))
> The pool is created each time when a UnionRDD is created , but the pool is not getting
shutdown. This is leaking thread/mem every slide interval of the window. 
> Easily reproducible with the below code. Just keep a watch on the number of threads.

> {code}
>     val sparkConf = new SparkConf().setMaster("local[*]").setAppName("TestLeak")
>     val ssc = new StreamingContext(sparkConf, Seconds(1))
>     ssc.checkpoint("checkpoint")
>     val rdd = ssc.sparkContext.parallelize(List(1,2,3))
>     val constStream = new ConstantInputDStream[Int](ssc,rdd)
>     constStream.window(Seconds(20),Seconds(1)).print()
>     ssc.start()
>     ssc.awaitTermination();
> {code}
> This happens only when the number of rdds to be unioned is above the value spark.rdd.parallelListingThreshold
(By default 10)
> Currently i'm working around by setting this threshold be a higher value. 
>  



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org


Mime
View raw message