[
https://issues.apache.org/jira/browse/YARN-8013?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16423280#comment-16423280
]
Weiwei Yang commented on YARN-8013:
-----------------------------------
Hi [~kkaranasos]
{quote}I would call the class {{TargetApplicationsNamespace}}. How does this
sound?
{quote}
Yes, I think this is a better name. I have renamed the class to this name in v6
patch.
{quote}It is small enough to be folded into the evaluate to be honest, and I
don't think the functionality of this class will be extended. Makes sense?
{quote}
Are you suggesting to remove this class? I prefer not to. Since I have renamed
the class to \{{TargetApplicationsNamespace}}, it looks pretty straightforward
to me to let the class be evaluable by a \{{TargetApplications}}
{code:java}
public abstract class TargetApplicationsNamespace implements
Evaluable<TargetApplications>
{code}
and \{{TargetApplications}} is used in a few places in UT as well. Second, this
way we keep our interface clean in case we need to extend what a namespace
needs to be eval against. For example, what if we want to support target types
by app? That will end up modifying the eval contract in
\{{TargetApplicationsNamespace}}. Does it make sense?
Uploaded v6 patch.
Thank you.
> Support APP-TAG namespace for allocation tags
> ---------------------------------------------
>
> Key: YARN-8013
> URL: https://issues.apache.org/jira/browse/YARN-8013
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Weiwei Yang
> Assignee: Weiwei Yang
> Priority: Major
> Attachments: YARN-8013.001.patch, YARN-8013.002.patch,
> YARN-8013.003.patch, YARN-8013.004.patch, YARN-8013.005.patch
>
>
> YARN-1461 adds *Application Tag* concept to Yarn applications, user is able
> to annotate application with multiple tags to classify apps. We can leverage
> this to represent a namespace for a certain group of apps. So instead of
> calling *app-label*, propose to call it *app-tag*.
> A typical use case is,
> There are a lot of TF jobs running on Yarn, and some of them are consuming
> resources heavily. So we want to limit number of PS on each node for such BIG
> players but ignore those SMALL ones. To achieve this, we can do following
> steps:
> # Add application tag "big-tf" to these big TF jobs
> # For each PS request, we add "ps" source tag and map it to constraint
> "{color:#d04437}notin, node, tensorflow/ps{color}" or
> "{color:#d04437}cardinality, node, tensorflow/ps{color}{color:#d04437}, 0,
> 2{color}" for finer grained controls.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
