Hadoop, “Browse the filesystem” broken link on ec2

问题

I am newbie to setting hadoop on EC2. I am trying to set-up single node hadoop setup. I have successfully installed and ran Hadoop 1.1.2 on one of the Amazon EC2 instance. All nodes, namenode, datanode, jobtracker, mapred node are on the same machine.

My core-site cofiguration is

<property>
    <name>fs.default.name</name>
    <value>hdfs://localhost:9000</value>
</property>

When I start the Hadoop, and go to web ui, namenode UI opens without any problem.

http://<namenode>:50070/dfshealth.jsp

But when clicked on the "Browse the filesystem", it redirects to

http://localhost:50075

which is incorrect. It should be something

http://<namenode>:50075. 

But in this case it failes.

Please help me to figure out the problem.

回答1:

I would advise using the fully qualified host names in your Hadoop configs, I've had issues when using localhost in the past. Namely, try setting the following (replace $hostname with yours):

In core-site.xml:

<property>
    <name>fs.default.name</name>
    <value>hdfs://$hostname.ec2.demdex.com:9000</value>
</property>

In hdfs-site.xml:

<property>
    <name>dfs.http.address</name>
    <value>$hostname.ec2.demdex.com:50070</value>
</property>

<property>
    <name>dfs.datanode.http.address</name>
    <value>$hostname.ec2.demdex.com:50075</value>
</property>

来源：https://stackoverflow.com/questions/16894715/hadoop-browse-the-filesystem-broken-link-on-ec2