[
https://issues.apache.org/jira/browse/YARN-6927?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16216080#comment-16216080
]
Daniel Templeton commented on YARN-6927:
----------------------------------------
Thanks for the updated patch, [~grepas]. At code level, it looks good. I have
a fewer higher level comments, though.
# Looks like CPU and memory are called "cpu-vcores" and "memory". To be more
consistent with the rest of resource types, it would be better to have "vcores"
and accept both "memory" and "memory-mb". (Did we already talk about this
before?)
# If I submit with -Dmapreduce.map.resource.memory=2048, it's treated as 2048B
and set up to 1024MB as the minimum. Ideally if the user doesn't give units,
we should infer them from the unit for the resource type. In this case, 2048
should be taken to be 2048MB. Same is true for the AM resources.
> Add support for individual resource types requests in MapReduce
> ---------------------------------------------------------------
>
> Key: YARN-6927
> URL: https://issues.apache.org/jira/browse/YARN-6927
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: resourcemanager
> Reporter: Daniel Templeton
> Assignee: Gergo Repas
> Attachments: YARN-6927.000.patch, YARN-6927.001.patch,
> YARN-6927.002.patch, YARN-6927.003.patch
>
>
> YARN-6504 adds support for resource profiles in MapReduce jobs, but resource
> profiles don't give users much flexibility in their resource requests. To
> satisfy users' needs, MapReduce should also allow users to specify arbitrary
> resource requests.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
