You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@ignite.apache.org by "Vladimir Ozerov (JIRA)" <ji...@apache.org> on 2019/02/14 13:30:00 UTC

[jira] [Comment Edited] (IGNITE-3203) Hadoop: permgen leak due to new statistics thread in Hadoop.

    [ https://issues.apache.org/jira/browse/IGNITE-3203?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16768201#comment-16768201 ] 

Vladimir Ozerov edited comment on IGNITE-3203 at 2/14/19 1:29 PM:
------------------------------------------------------------------

Not relevant at the moment.


was (Author: vozerov):
https://issues.apache.org/jira/browse/IGNITE-3187

> Hadoop: permgen leak due to new statistics thread in Hadoop.
> ------------------------------------------------------------
>
>                 Key: IGNITE-3203
>                 URL: https://issues.apache.org/jira/browse/IGNITE-3203
>             Project: Ignite
>          Issue Type: Task
>          Components: hadoop
>    Affects Versions: 1.5.0.final, 1.6
>            Reporter: Vladimir Ozerov
>            Priority: Major
>
> *Problem*
> HADOOP-12829 ticket was implemented recently. It introduces special worker thread which monitors file system statistics. When several {{HadoopClassLoader}}s are created, one thread start for each of them. As this thread is never stopped, classloaders are never unloaded -> leak.
> *Solution*
> We need to investigate on how to hack into this process and prevent start of these thread. May be we will have to implement our own threads serving the same purpose.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)