You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Josh Rosen (JIRA)" <ji...@apache.org> on 2014/10/26 08:50:33 UTC
[jira] [Commented] (SPARK-4090) Memory leak in snappy-java
1.1.1.4/5
[ https://issues.apache.org/jira/browse/SPARK-4090?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14184423#comment-14184423 ]
Josh Rosen commented on SPARK-4090:
-----------------------------------
I rolled back earlier today, so the build should be fixed now.
> Memory leak in snappy-java 1.1.1.4/5
> ------------------------------------
>
> Key: SPARK-4090
> URL: https://issues.apache.org/jira/browse/SPARK-4090
> Project: Spark
> Issue Type: Bug
> Components: Spark Core
> Affects Versions: 1.2.0
> Reporter: Davies Liu
> Priority: Blocker
> Attachments: screenshot-12.png
>
>
> There is memory-leak bug in snappy-java 1.1.1.4/5, we should rollback to 1.1.1.3 or wait for bugfix.
> The jenkins tests timeouted or OOM multiple times recently. While test it locally, I got the heap dump of leaked JVM:
> Then found that it's a bug in recent releases of snappy-java:
> {code}
> + inputBuffer = inputBufferAllocator.allocate(inputSize);
> + outputBuffer = inputBufferAllocator.allocate(outputSize);
> {code}
> The outputBuffer is allocated from inputBufferAllocator but released to outputBufferAllocator: https://github.com/xerial/snappy-java/issues/91
--
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