Hadoop on Local FileSystem

Question

I'm running Hadoop on a pseudo-distributed. I want to read and write from Local filesystem by Abstracting the HDFS for my job. Am using the file:/// parameter. I followed this link.

This is the file contents of core-site.xml,

<configuration>
 <property>
  <name>hadoop.tmp.dir</name>
  <value> /home/abimanyu/temp</value>   
 </property>

 <property>
  <name>fs.default.name</name>
  <value>hdfs://localhost:54310</value>
 </property>

</configuration>

This is the file contents of mapred-site.xml,

<configuration>

 <property>
  <name>mapred.job.tracker</name>
  <value>localhost:54311</value>
 </property>

 <property>
    <name>fs.default.name</name>
    <value>file:///</value>
 </property>

 <property>
    <name>mapred.tasktracker.map.tasks.maximum</name>
    <value>1</value>
 </property>

 <property>
    <name>mapred.tasktracker.reduce.tasks.maximum</name>
    <value>1</value>
 </property>

</configuration>

This is the file contents of hdfs-site.xml,

<configuration>

 <property>
  <name>dfs.replication</name>
  <value>1</value>
 </property>
</configuration>

This is the error I get when I try to start the demons(using start-dfs or start-all),

localhost: Exception in thread "main" java.lang.IllegalArgumentException: Does not contain a valid host:port authority: file:///
localhost:      at org.apache.hadoop.net.NetUtils.createSocketAddr(NetUtils.java:164)
localhost:      at org.apache.hadoop.hdfs.server.namenode.NameNode.getAddress(NameNode.java:212)
localhost:      at org.apache.hadoop.hdfs.server.namenode.NameNode.getAddress(NameNode.java:244)
localhost:      at org.apache.hadoop.hdfs.server.namenode.NameNode.getServiceAddress(NameNode.java:236)
localhost:      at org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode.initialize(SecondaryNameNode.java:194)
localhost:      at org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode.<init>(SecondaryNameNode.java:150)
localhost:      at org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode.main(SecondaryNameNode.java:676)

What is strange to me is that this reading from local file system works completely fine in hadoop-0.20.2 but not in hadoop-1.2.1. Has anything changed from initial release to the later version ? Let me know how to read from Local File system for a Hadoop JAR.

Solution

You can remove the fs.default.name value from your mapred-site.xml file - this should only be in the core-site.xml file.

If you want to run on your local file system, in a pseudo mode, this is typically achieved by running in what's called local mode - by setting the fs.default.name value in core-site.xml to file:/// (you currently have it configured for hdfs://localhost:54310).

The stack trace you are seeing is when the secondary name node is starting up - this isn't needed when running in 'local mode' as there is no fsimage or edits file for the 2NN to work against.

Fix up your core-site.xml and mapred-site.xml. Stop all hadoop daemons and just start the map-reduce daemons (Job Tracker and Task Tracker).