hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tsuyoshi OZAWA (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-2324) Race condition in continuousScheduling for FairScheduler
Date Sun, 20 Jul 2014 11:28:38 GMT

    [ https://issues.apache.org/jira/browse/YARN-2324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14067889#comment-14067889
] 

Tsuyoshi OZAWA commented on YARN-2324:
--------------------------------------

[~zxu], thank you for reporting and your contribution! As [~ywskycn] mentioned, we're addressing
the problem on YARN-2273. Let's close this ticket as duplicated.

> Race condition in continuousScheduling for FairScheduler
> --------------------------------------------------------
>
>                 Key: YARN-2324
>                 URL: https://issues.apache.org/jira/browse/YARN-2324
>             Project: Hadoop YARN
>          Issue Type: Bug
>            Reporter: zhihai xu
>            Assignee: zhihai xu
>         Attachments: YARN-2324.000.patch
>
>
> Race condition in continuousScheduling for FairScheduler.
> removeNode can run when continuousScheduling is called in schedulingThread. If the node
is removed from nodes, nodes.get(n2) and getFSSchedulerNode(nodeId) will be null. So we need
add lock to remove the NPE/race conditions.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message