You are viewing a plain text version of this content. The canonical link for it is here.

Posted to yarn-issues@hadoop.apache.org by "Junping Du (JIRA)" <ji...@apache.org> on 2015/02/20 17:08:12 UTC

[jira] [Commented] (YARN-3115) [Aggregator wireup] Work-preserving restarting of per-node aggregator

    [ https://issues.apache.org/jira/browse/YARN-3115?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14329103#comment-14329103 ] 

Junping Du commented on YARN-3115:
----------------------------------

Hi [~zjshen], do you already start the work on this? If not, I am quite familiar with NM work preserving in restart. Can I take this JIRA on? Thanks!

> [Aggregator wireup] Work-preserving restarting of per-node aggregator
> ---------------------------------------------------------------------
>
>                 Key: YARN-3115
>                 URL: https://issues.apache.org/jira/browse/YARN-3115
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: timelineserver
>            Reporter: Zhijie Shen
>            Assignee: Zhijie Shen
>
> YARN-3030 makes the per-node aggregator work as the aux service of a NM. It contains the states of the per-app aggregators corresponding to the running AM containers on this NM. While NM is restarted in work-preserving mode, this information of per-node aggregator needs to be carried on over restarting too.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)