[
https://issues.apache.org/jira/browse/SPARK-19644?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16234638#comment-16234638
]
Shixiong Zhu commented on SPARK-19644:
--------------------------------------
I happened to investigate a similar issue and found out the leak is caused by
Scala reflection. Please see my comment in
https://github.com/scala/bug/issues/8302
My workaround is calling
"scala.reflect.runtime.universe.asInstanceOf[scala.reflect.runtime.JavaUniverse].undoLog.clear()"
manually to clean up these garbage objects. You need to put this line in the
same thread that you create Dataset/DataFrame as the leak happens in a thread
local object. I think the best place in Spark streaming is foreachRDD.
> Memory leak in Spark Streaming
> ------------------------------
>
> Key: SPARK-19644
> URL: https://issues.apache.org/jira/browse/SPARK-19644
> Project: Spark
> Issue Type: Bug
> Components: DStreams
> Affects Versions: 2.0.2
> Environment: 3 AWS EC2 c3.xLarge
> Number of cores - 3
> Number of executors 3
> Memory to each executor 2GB
> Reporter: Deenbandhu Agarwal
> Priority: Major
> Labels: memory_leak, performance
> Attachments: Dominator_tree.png, Path2GCRoot.png, heapdump.png
>
>
> I am using streaming on the production for some aggregation and fetching data
> from cassandra and saving data back to cassandra.
> I see a gradual increase in old generation heap capacity from 1161216 Bytes
> to 1397760 Bytes over a period of six hours.
> After 50 hours of processing instances of class
> scala.collection.immutable.$colon$colon incresed to 12,811,793 which is a
> huge number.
> I think this is a clear case of memory leak
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
