Explain to me the correct way to get a Hadoop FileSystem object so that I can use it for reading from HDFS as well as writing to HDFS

Question

Can anyone help me with the right path to get a Hadoop file system so that I can read and write a certain file in HDFS? I tried a few methods as follows:

final Configuration conf = new Configuration();
conf.addResource(new Path("/usr/local/hadoop/etc/hadoop/core-site.xml"));
conf.addResource(new Path("/usr/local/hadoop/etc/hadoop/hdfs-site.xml"));

final FileSystem fs = FileSystem.get(conf);

The Class configuration documentation says the core-site.xml properties automatically get loaded when an object is created and the required file is in the classpath and there would be no need to set it again. 

The HDFS-site.xml is not added, yet it seems to work fine.

Is it a recommended option to just configure Core-site.xml and skip HDFS-site.xml or is it mandatory to define both? please let me know.

Answer 1

FileSystem needs only one configuration key to successfully connect to HDFS. Previously it was fs.default.name. From yarn onward, it's changed to fs.defaultFS. So the following snippet is sufficient for the connection.

Configuration conf = new Configuration();
conf.set(key, "hdfs://host:port");  // where key="fs.default.name"|"fs.defaultFS"

FileSystem fs = FileSystem.get(conf);       

I would suggest checking the core-site.xml which key exists. Set the same value associated with it in conf. If the machine from where you are running the code doesn't have the hostname mapping, put its IP. In MapR cluster value will have a prefix like MapRfs://.