[
https://issues.apache.org/jira/browse/YARN-5532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15436943#comment-15436943
]
Arun Suresh commented on YARN-5532:
-----------------------------------
bq. I think ResourceLocalizationService should completely abstract away the
requester and just manage everything via a RequesterID - resource usage per
requester
I agree. But I guess we have to think about how this affects the lifetime of
the downloaded Resource now. Earlier, since localization was tied to container
lifecycle, when a container completes, we know exactly which resources need to
be cleaned up. We would now probably need some form of reference checking /
garbage collection.
Would be interesting to see how we can plug in YARN-4597 too. I am guessing all
we need is to send the container lifecycle trigger to the localizing service.
But we can discuss more during implementation.
> Make ResourceLocalizationService inside NodeManager independent of containers
> -----------------------------------------------------------------------------
>
> Key: YARN-5532
> URL: https://issues.apache.org/jira/browse/YARN-5532
> Project: Hadoop YARN
> Issue Type: New Feature
> Reporter: Jian He
> Assignee: Jian He
>
> The goal is to enable NodeManagers to localize resources while container is
> running. Today, resource-localization is always the first step before
> starting a container. It will be useful if YARN can localize the resources
> continuously even while container is running.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
