[
https://issues.apache.org/jira/browse/TEZ-3643?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15890557#comment-15890557
]
Siddharth Seth commented on TEZ-3643:
-------------------------------------
Thanks for the review. Committing.
In terms of 0.7 - I don't think we should be porting a lot of changes to the
branch, and would prefer not backporting this. This was mainly seen as a result
of llap creating and discarding containers really fast. For regular YARN
containers, this is unlikely to be hit (there may be some degraded performance
if an AM lives for a really long time - depends on the AM size.) If you feel
strongly about back porting it - I'll post a patch for 0.7.
> Long running AMs can go out of memory due to retained AMContainer instances
> ---------------------------------------------------------------------------
>
> Key: TEZ-3643
> URL: https://issues.apache.org/jira/browse/TEZ-3643
> Project: Apache Tez
> Issue Type: Bug
> Reporter: Siddharth Seth
> Assignee: Siddharth Seth
> Priority: Critical
> Attachments: TEZ-3643.01.patch
>
>
> 3.4million instances took over a 3GB heap.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
