CODEWITHSUNDEEP
hadoopbigdataclouderahortonworks-data-platform
Microsoft Developer
Microsoft Developer

Reputation: 5459

Create or get environment for Hadoop practice

I have started reading about hadoop. I want to learn practicals as well. As hadoop is distributed environment and made to run on linux, I cannot practice it on my local machine which runs windows. Is there any cloud or virtual machines available on internet where I can set up entire environment by my own and start learning? It is also fine I can get pre-configured hadoop cluster but I will prefer to configure by my own.

Upvotes: 0

Views: 3192

Answers (3)

gszecsenyi
gszecsenyi

Reputation: 93

Cloudera ( cloudera.com) has some preconfigured one-node environments. It is only 3GB, but you find every BigData components, what you need.

http://www.cloudera.com/downloads/quickstart_vms.html

Oracle has also a VM, free for development nad testing: Oracle BigData Lite

http://www.oracle.com/technetwork/database/bigdata-appliance/oracle-bigdatalite-2104726.html

and Hortonworks has also a preconfigured sandbox.

I prefer the Cloudera environment. Oracle is good, if yoou wish to test the Sqoop component.

Upvotes: 3

Baeumla
Baeumla

Reputation: 473

I also wanted to learn something about Hadoop, so I bought a Raspberry Pi, installed Raspbian wheezy and Hadoop following these instructions.

It's very well explained how to set up a Single-Node Cluster. The Hadoop on a Raspberry Pi is not very fast, but in my opinion it's fast enough to learn the basics of Hadoop.

If you want to set up a single-node cluster on your local machine, I would prefer to use a distribution, as user3341955 already mentioned. I took Hortonworks because there was a very simple setup-guide for Virtualbox on Mac OS X.

And the setup-guide for VMware Player on Windows:

Have fun with Hadoop :)

Upvotes: 0

Sean Owen
Sean Owen

Reputation: 66891

Since you are looking for a cloud-based environment to play around, I suggest the Cloudera Live demo, which is really the Hue demo. It's basically a real instance of Hue (which is the UI that non-admin users would use and see) in front of a real cluster that is open to the public. You can browse the file system, issues queries, etc.

If you want to explore in more than superficial detail, yes you should run your own toy cluster. If you like the Cloudera distro, note that Cloudera Manager is free and is the installer program as well that can turn up a cluster in the cloud, like on Amazon EC2, for you.

Upvotes: 0

Related Questions