I have been trying to find a way to provide a clear working environment to my class participants on Hadoop as I have had few hiccups in my previous weekend class at Cochin. My Cochin class photograph with participants shows the geeks with passion to learn about this Hadoop but with disappointments when they found that their laptop is NOT suiting the environment.
I have been using the VM provided from Cloudera. This is easier as it had a pre-installation of Hadoop along with the utilities such as pig, hive, etc. But what challenged me what that we had asked participants to bring the laptop for the practical classes and everyone had brought different systems in terms of hardware and operating systems. Few challenges were
1. Cloudera image was built for 32 bit and 64 bit systems had issues in running it.
2. Some of the laptops participants brought were pretty old that they don't have virtualisation option in the BIOS to be enabled
3. Some had their BIOS locked so could not enable the VT option
4. Some had issues in the Linux boot after VM started resulting in Kernel panic.
I am not going to speak here about how I went and fixed them at that point. But I am wondering an easy way out to resolve these kind of issues for my future classes.
Initially I was planning on to give USB boot linux so the participants can work on the exercises booting their system from Linux USB drive but I was unsuccessful in having the live persistent for the OS and it required another USB drive to be formatted and installed as a full install to run the complete set of exercises so I gave this one up.
Now what I have been thinking is that I will use my VMWare workstation to create a new Linux image from the ISO file and install all the necessary packages in it and prepare a clone to distribute it to my participants. I hope it works but for all those who have similar thoughts, here is a quick guideline of what I did
1. Install VMware workstation if you have not done that. This is the starting point.
2. Once you have the VMware workstation, you will also need a Linux ISO image. I have used Ubuntu 12.04 LTS Precise Pangolin for my plan. Download the ISO if you have not got it from www.ubuntu.com
3. Open the VMware workstation and create a new virtual machine from the ISO file. There are lots of materials on-line to see how you do it so I am not getting to this detail here.I used a typical install with 2 GB RAM and 20 GB hard disk in my configuration. Note ensure that you create the new VM without powering it ON after the creation.
4. Now power ON the VM with the configuration. I had issues with the VMware tools installation as part of the Easy install of Ubuntu. So cancel the VMware install and reboot the Linux operating system in the VM.
5. Wait until the Linux boots up and you are in the guest completely as below.
6. Now click the install VMware tools on the VM menu while being inside the Linux operating system and you will see the VMware tools mounted
7. Now install the VMware tools by copying the tar file to your home drive and execute the install script. Refer on-line VMware help for details.
8. After this I copied Oracle Java and Hadoop to the home folder. And I followed the link below to configure Hadoop.
9. Now that I have Hadoop installed and working on my VMware Linux, I powered off the VM. And I created a Full clone of the VM. I have zipped this along with all the supplementary files that VMware created.
I am going to distribute this zip file to the participants of my class so they run everything from the VM clone I had created for Hadoop. I will write more about this if I see any issues here.
0 Comments:
Post a Comment