Hadoop on your PC: Cloudera's CDH4 virtual machine

If you're not running the full-fledged VMWare product, never fear.  Just download the free VMWare player, then use it to run the CDH4 image.

If you put the VM on your network, you'll be able to use Hadoop from your own PC (I'll show you how in the next two screens).  To make this work, just get into VMWare's Virtual Machine Settings dialog box, select the Network Adapter device and then the Connected check box.

Now you're ready to get the VM's IP address and surf to it from your own browser.  Continue to the next screen for details on getting the IP address.

Cloudera's CDH4 runs its own Web server and a Web-based user interface, called Hue, sporting consoles for MapReduce, HDFS and Hive, along with browser-based command line shells for HBase and Pig.

You can run Hue from your own PC's (i.e. the host's) Web browser, as long as you know the IP address of the VM.  Inside the VM, open an X terminal emulator command prompt (by clicking the highlighted icon at the bottom of the screen) and enter the ifconfig command to determine the address.  You'll see what to do with that address in the next screen...

To get to Hue, just open a browser on your own host PC (use Firefox or Chrome -- I had trouble with IE), and navigate to port 8888 at the IP address you just discovered.  For example, if the IP address were 192.168.1.35 (as shown in the previous screen), you'd type the following into your browser's address bar:
http://192.168.1.35:8888

This will take you to the Hue login screen, pictured above.  Login using "cloudera" as both the user name and password.

After logging in to Hue, its browser-based user interface for Hive, called Beeswax, comes up by default.  Beeswax allows you to enter SQL (well, technically, HiveQL) queries and see the results.  You can also save queries, upload files, edit Hive's settings, or create user-defined functions.

Used to a file system GUI?  Don't feel like typing commands like "lsr" and "mkdir" at the shell prompt?  Never fear, because Hue's browser-based HDFS user interface is here.  Just click the file cabinet icon in Hue's toolbar to get there.

Shown above is Hive's root folder.  Clicking on the "warehouse" link would let you view the files which are queryable as tables in Hive.

Click the Job Designer icon (4th from the left in Hue's toolbar), then the Create Mapreduce Design button, and up comes a screen that lets you run a jar-based MapReduce job.  Give the job a name and a description, specify the jar file, add properties (parameters) and property values, then click Save.

You'll be automatically redirected to the Job Designs screen where you can click the Submit button for the job design you just built.  You can also click the Create Streaming Design button to design a job whsoe mapper and reducer code is written in a language other than Java.

Click the Hue Shell Toolbar button to get to browser-based command line shells for Pig and HBase.  Below the toolbar, you'll find one clickable link for each.  The Pig shell (called "Grunt") is shown here.  At this prompt you can enter Pig Latin commands interactively or run .pig scripts.

Click the HBase Shell link beneath Hue's toolbar and you'll come to HBase's command line interface.  Read more about the commands you can use in this interface here.

---

That's pretty much all there is to working with the CDH 4 VM!  Of course, if you prefer the command line for everything, then you can go back to an X terminal window and do all your work from there.

No matter which way you work, you can now start to learn Hadoop, Pig, Hive and more.  And you can build your own clusters later on, if you're into that sort of thing :-)