[
https://issues.apache.org/jira/browse/SPARK-12906?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15106548#comment-15106548
]
Sasi commented on SPARK-12906:
------------------------------
Hi,
I'm running endless script that does query on my table using SqlSpark.
I did a dump using jmap tool before the test and after the test.
I also run GC on the environment and I still saw the size of LognSQLMetricValue
didn't change and only increased.
My suspect is on the count method, because i'm running
dataFrame.distinct().count().
Sasi
> LongSQLMetricValue cause memory leak on Spark 1.5.1
> ---------------------------------------------------
>
> Key: SPARK-12906
> URL: https://issues.apache.org/jira/browse/SPARK-12906
> Project: Spark
> Issue Type: Bug
> Affects Versions: 1.5.1
> Reporter: Sasi
> Attachments: screenshot-1.png
>
>
> Hi,
> I was upgrade my spark from 1.5.0 to 1.5.1 after saw that the
> scala.util.parsing.combinator.Parser$$anon$3 cause memory leak.
> Now, after doing another dump heap I notice, after 2 hours, that
> LongSQLMetricValue cause memory leak.
> Didn't see any bug or document about it.
> Thanks,
> Sasi
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
