[
https://issues.apache.org/jira/browse/YARN-417?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13595409#comment-13595409
]
Sandy Ryza commented on YARN-417:
---------------------------------
Bikas,
Ok, I will join with the handler thread on stop and we can just throw an
exception to fail early when it's called by the handler thread.
Hitesh, this is how it works currently. Whenever a heartbeat returns,
AMRMClient stores its info about the cluster internally and makes it available
through getNumClusterNodes and getClusterAvailableResources. AsyncAMRMClient
exposes this through wrappers with the same name. Is this insufficient? It's
hard for me to imagine an AM that would need to take an action on a change in
the cluster's available resources, when there aren't any other callbacks. Does
anybody have any scenarios in mind?
As discussed earlier, I'm not able to think of any scenarios when an AM would
need to know what containers were allocated before what containers completed,
and having separate callbacks for these seems a little simpler to me. Having
both in the same callback requires the user to understand that the method will
be called if one of the arguments is empty, but not if both are. That said,
I'm not strongly attached either way, so if there's consensus on merging them,
I'm happy to make that change.
> Add a poller that allows the AM to receive notifications when it is assigned
> containers
> ---------------------------------------------------------------------------------------
>
> Key: YARN-417
> URL: https://issues.apache.org/jira/browse/YARN-417
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: api, applications
> Affects Versions: 2.0.3-alpha
> Reporter: Sandy Ryza
> Assignee: Sandy Ryza
> Attachments: AMRMClientAsync-1.java, AMRMClientAsync.java,
> YARN-417-1.patch, YARN-417-2.patch, YARN-417-3.patch, YARN-417.patch,
> YarnAppMaster.java, YarnAppMasterListener.java
>
>
> Writing AMs would be easier for some if they did not have to handle
> heartbeating to the RM on their own.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
