RE: Slow mapreduce using Hbase , regardless on number of machines

[Yair Even-Zohar]

I believe you have to set the HADOOP_CONF_DIR to point to the conf
directory but I'm not certain here.
I just copied the hadoop-site.xml to my hbase conf directory and that
solved the problem

-Yair

-----Original Message-----
From: Erik Holstad [mailto:[EMAIL PROTECTED] 
Sent: Monday, July 14, 2008 5:36 AM
To: hbase-user@hadoop.apache.org
Subject: Re: Slow mapreduce using Hbase , regardless on number of
machines

Aha, yes I see that it is running in local mode

, but how do you change that.
I have setup the paths in
hadoop-site:
<configuration>
    <property>
        <name>fs.default.name</name>
        <value>mb0:9000</value>
    </property>
    <property>
        <name>mapred.job.tracker</name>
        <value>mb0:9001</value>
    </property>
    <property>
        <name>dfs.replication</name>
        <value>3</value>
    </property>
</configuration>


and uncommented export
HADOOP_CLASSPATH=/home/hbase/jerl:/home/hbase/hbase/hbase-0.1.3-NOJ-3.ja
r:/home/hbase/hbass/conf
in the hadoop-env file.

Regards Erik