[
https://issues.apache.org/jira/browse/YARN-6384?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16206295#comment-16206295
]
Miklos Szegedi commented on YARN-6384:
--------------------------------------
Thank you for the patch [~lolee_k],
DEFAULT_NM_LINUX_CONTAINER_CGROUPS_MAX_RESOURCE_USAGE_PERCENT should be 1.0 by
default I think not to break backward compatibility and issues are easier to
debug in a default cluster.
Please fix the outstanding checkstyle issues.
Just a note on the overall design. I think the patch does it correctly, it
allows a higher limit to be set in case of strict mode, not affecting
non-strict mode. In a future jira we may consider mixing the two, allowing
weights to be set with a strict limit. However, that would be a breaking
change, since strict and non-strict mode are not distinct anymore.
> Add configuratin to set max cpu usage when strict-resource-usage is false
> with cgroups
> --------------------------------------------------------------------------------------
>
> Key: YARN-6384
> URL: https://issues.apache.org/jira/browse/YARN-6384
> Project: Hadoop YARN
> Issue Type: Improvement
> Reporter: dengkai
> Attachments: YARN-6384-0.patch, YARN-6384-1.patch, YARN-6384-2.patch,
> YARN-6384-3.patch
>
>
> When using cgroups on yarn, if
> yarn.nodemanager.linux-container-executor.cgroups.strict-resource-usage is
> false, user may get very more cpu time than expected based on the vcores.
> There should be a upper limit even resource-usage is not strict, just like a
> percentage which user can get more than promised by vcores. I think it's
> important in a shared cluster.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
