[
https://issues.apache.org/jira/browse/YARN-5531?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15999045#comment-15999045
]
Subru Krishnan commented on YARN-5531:
--------------------------------------
Thanks [~botong] for addressing my comments. I looked at the latest patch is
pretty close, have a few comments:
* The pom changes for *hadoop-yarn-server-common* test dependencies are
already committed as part of YARN-5411, can you kindly rebase.
* Don't remove {{FederationProxyProviderUtil}}, we need it to manage
Federation dependencies which currently has leaked into
{{YarnServerSecurityUtils}}. Maybe {{AMRMProxyServiceUtils}} is a better place
to put the *createRMProxy* but either ways the general utility class should not
have any dependency to Federation.
* bq. hreadpool is not a Yarn service, and no serviceStart call
I don't get your comment as the threadpool is managed by
{{UnmanagedAMPoolManager}} which is indeed a service:
{code} public class UnmanagedAMPoolManager extends AbstractService {code}
* {{UnmanagedAMLauncher::submitAppAndGetAppId}} can be void.
* {code} if (!appAttemptId.equals(this.attemptId)) { {code} - I feel this
check is not correct as once we have HA (YARN-6128), we will have multiple
attempts in home SC but will always have only one for secondary as those are
UAMs.
* Nit: In {{TestUnmanagedApplicationManager::testForceKill}}, is there any
specific exception message that you can assert as the second call can potential
fail for other reasons.
> UnmanagedAM pool manager for federating application across clusters
> -------------------------------------------------------------------
>
> Key: YARN-5531
> URL: https://issues.apache.org/jira/browse/YARN-5531
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: nodemanager, resourcemanager
> Reporter: Subru Krishnan
> Assignee: Botong Huang
> Attachments: YARN-5531-YARN-2915.v1.patch,
> YARN-5531-YARN-2915.v2.patch, YARN-5531-YARN-2915.v3.patch,
> YARN-5531-YARN-2915.v4.patch, YARN-5531-YARN-2915.v5.patch,
> YARN-5531-YARN-2915.v6.patch
>
>
> One of the main tenets the YARN Federation is to *transparently* scale
> applications across multiple clusters. This is achieved by running UAMs on
> behalf of the application on other clusters. This JIRA tracks the addition of
> a UnmanagedAM pool manager for federating application across clusters which
> will be used the FederationInterceptor (YARN-3666) which is part of the
> AMRMProxy pipeline introduced in YARN-2884.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
