Managing hdfs in pseudo-distributed hadoop mode

开发者_Go百科I want to do some computation with hadoop and mahout on my quad core machine, so I am using hadoop in pseudo-distributed mode.

The problem is that the space on my root drve is limited, so how can I configure it to use space available on some other external hard drive.

---

You can configure where hdfs strores its data. Add the following to your conf/hdfs-site.xml:

<property>
    <name>dfs.data.dir</name>
    <value>__path_to_where_you_want_to_store_your_data/hdfs/data/</value>
</property>
<property>
    <name>dfs.name.dir</name>
    <value>__path_to_where_you_want_to_store_your_data/hdfs/name/</value>
</property>

After theese changes you will have to format your namenode:

hadoop namenode -format