hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Vladimir Klimontovich <klimontov...@gmail.com>
Subject Re: How running hadoop without command line?
Date Thu, 27 Aug 2009 14:51:18 GMT
There is a class called JobClient. You can run jobs using these class  
from java app.
Also hadoop jar adds $HADOOP_HOME/conf/hadoop-site.xml and  
$HADOOP_HOME/conf/hadoop-conf.xml
xml to classpath, so JobClient already knows which jobtracker and FS  
should be used for running map process.

So when you are running hadoop jobs inside other application you  
should take care of setting valid properties
for jobtracker address, FS address and etc. You can set it manually or  
use Configuratio.addResource for
adding $HADOOP_HOME/conf/hadoop-conf.xml and $HADOOP_HOME/conf/hadoop- 
site.xml .

On Aug 27, 2009, at 1:06 PM, radar.sxl wrote:

>
> When run a hadoop project jar, We'll use
> $ bin/hadoop jar ***.jar
> but if I hava a java app, i want to run mapreduce job inside it use  
> hadoop
> cluster, is three anyway?
> -- 
> View this message in context: http://www.nabble.com/How-running-hadoop-without-command-line--tp25167322p25167322.html
> Sent from the Hadoop core-user mailing list archive at Nabble.com.
>

---
Vladimir Klimontovich,
skype: klimontovich
GoogleTalk/Jabber: klimontovich@gmail.com
Cell phone: +7926 890 2349


Mime
View raw message