You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@bigtop.apache.org by "Evans Ye (JIRA)" <ji...@apache.org> on 2017/01/07 15:26:58 UTC

[jira] [Commented] (BIGTOP-2663) puppet hadoop module: Consolidate memory resource settings

    [ https://issues.apache.org/jira/browse/BIGTOP-2663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15807657#comment-15807657 ] 

Evans Ye commented on BIGTOP-2663:
----------------------------------

Sounds convincing for *yarn.nodemanager.vmem-pmem-ratio* to me.
And the rest if the code looks good.
+1

> puppet hadoop module: Consolidate memory resource settings 
> -----------------------------------------------------------
>
>                 Key: BIGTOP-2663
>                 URL: https://issues.apache.org/jira/browse/BIGTOP-2663
>             Project: Bigtop
>          Issue Type: Bug
>    Affects Versions: 1.1.0
>            Reporter: Olaf Flebbe
>            Assignee: Olaf Flebbe
>             Fix For: 1.2.0
>
>         Attachments: 0001-BIGTOP-2663-puppet-hadoop-module-Consolidate-memory-.patch
>
>
> The memory resource settings for hadoop are outdated.
> Now the settings in mapred-site.xml should be used
> {code}
> mapreduce.map.java.opts
> mapreduce.reduce.java.opts
> {code}
> These are set now to {{-Xmx1024m}} (This was hardcoded before)
> Additionally one can now optionally set the maxmimum (resident) memory
> for map and reduce jobs
> {code}
> mapreduce.map.memory.mb
> mapreduce.reduce.memory.mb
> {code}
> And last but not least, will set {{yarn.nodemanager.vmem-pmem-ratio}} to 100:
> There is the public misconception that virtual memory is a limiting resource.
> That's only the case for 32Bit Adress spaces, not anymore. 
> See for instance for http://stackoverflow.com/questions/561245/virtual-memory-usage-from-java-under-linux-too-much-memory-used
> for an rather up to date detailled explanation, why vmem doesn't matter.
> So we allow it to be tremendiously large. Why does it matter, anyhow? Java8 seems to use memory mapped I/O agressivly now, and the virtual memory in the hadoop mapred container became exhausted when the resident memory is only 15% used.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)