[
https://issues.apache.org/jira/browse/YARN-6927?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16205011#comment-16205011
]
Sunil G commented on YARN-6927:
-------------------------------
Yes. This sounds a better syntax {{(memory=x unit, vcore=y unit, gpu=z unit,
...)}}. With more resource-types getting added in YARN, defining arbitrary
resource names here will be inlining with the same resource name in
resource-types.xml. With YARN-7307, I assume that server and client will have
same synonymous meaning for resource types.
I am not very sure whether its a problem or not.
{{mapreduce.map/reduce.resource.<resource-vectors>}} here resource vectors are
{noformat}
memory
vcores
disk-space
{noformat}
looks like a very clean config model.
Could we consider other resource specifications like
{{yarn.app.mapreduce.am.resource.mb/cpu-vcores}} as a standard way to define
CPU and Memory is users wants to use, rather than an overwriting model? In case
when both are specified, i am more in favor of throwing an exception to end
user asking to specify in one format.
> 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-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]
