hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ted Yu (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-6695) Job Tracker should be able to refresh classpath without restarting
Date Sun, 11 Apr 2010 16:52:42 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-6695?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12855751#action_12855751
] 

Ted Yu commented on HADOOP-6695:
--------------------------------

One possibility is for JobClient to pass its effective CLASSPATH, e.g. through (a new) mapred.job.classpath
JobConf property, to JobTracker. JobTracker examines this classpath against its classpath
to find new jars and put them into DistributedCache.

The benefit is that when user adds new jars to HADOOP_CLASSPATH which get picked up by JobClient,
JobTracker is able to honor the addition.


> Job Tracker should be able to refresh classpath without restarting
> ------------------------------------------------------------------
>
>                 Key: HADOOP-6695
>                 URL: https://issues.apache.org/jira/browse/HADOOP-6695
>             Project: Hadoop Common
>          Issue Type: Improvement
>          Components: conf
>    Affects Versions: 0.20.1
>            Reporter: Ted Yu
>
> When I tried to run HBase export in cluster mode, I copied zookeeper-3.2.1.jar to hadoop
lib directory.
> But initial attempts failed with java.lang.ClassNotFoundException because job tracker
wasn't restarted.
> It is desirable for job tracker to pick up new classpath without restarting which is
sometimes hard to do in production environment.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message