hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sunil G (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-4024) YARN RM should avoid unnecessary resolving IP when NMs doing heartbeat
Date Thu, 06 Aug 2015 18:48:04 GMT

    [ https://issues.apache.org/jira/browse/YARN-4024?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14660602#comment-14660602
] 

Sunil G commented on YARN-4024:
-------------------------------

Thanks [~leftnoteasy] for sharing detailed case.

Yes, what I meant is a case where lookup fails and we need to resolve again. And further to
this then we ll end up in having black/white level issue. 
I feel there will not be much of a problem to do only caching provided we need to clear cache
from some external way once a threshold is reached. This is to save too much of invalid IPs
as in your case. Will it be fine?

> YARN RM should avoid unnecessary resolving IP when NMs doing heartbeat
> ----------------------------------------------------------------------
>
>                 Key: YARN-4024
>                 URL: https://issues.apache.org/jira/browse/YARN-4024
>             Project: Hadoop YARN
>          Issue Type: Bug
>            Reporter: Wangda Tan
>
> Currently, YARN RM NodesListManager will resolve IP address every time when node doing
heartbeat. When DNS server becomes slow, NM heartbeat will be blocked and cannot make progress.



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

Mime
View raw message