hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Harsh J (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (HADOOP-378) too many files are distributed to slaves nodes
Date Sat, 16 Jul 2011 16:47:59 GMT

     [ https://issues.apache.org/jira/browse/HADOOP-378?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Harsh J resolved HADOOP-378.
----------------------------

    Resolution: Incomplete

(This one might have gone stale over time, and perhaps does not apply anymore)

This one needs to explain more about what its talking about - deployment of Hadoop via HADOOP_MASTER
which uses rsync (and thereby does not pass out incrementals)?

This would be resolved by better packaging, which is gonna be addressed in future via Bigtop
and the like, hopefully. A package for examples, a package for test, a package for clients.
With mavenized builds, this is very easy to do and is on its way.

Resolving as Incomplete, needs more desc.

> too many files are distributed to slaves nodes
> ----------------------------------------------
>
>                 Key: HADOOP-378
>                 URL: https://issues.apache.org/jira/browse/HADOOP-378
>             Project: Hadoop Common
>          Issue Type: Bug
>          Components: conf
>            Reporter: Yoram Arnon
>            Priority: Trivial
>
> currently all the svn tree is sync'ed to the slaves on startup, excluding only .svn files.
> that includes the example jar file, sources, and other files that are not required.
> It would  be better to pick and choose the files that get sync'ed. It will improve sync
times on startup, and can prevent some name conflicts, since all the jar files are also included
in the classpath by the hadoop script.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message