Install Eclipse on Cloudera HDP Sand Box

221 views Asked by At

I was using earlier Cloudera QuickStart VM using oracle virtual box. This QuickStart VM provided the pre-installed Eclipse and web browser. So I can install the Intellij or any other software on this VM by the web browser. But Un-fortunately Cloudera stopped the QuickStart VM and they are providing HDP Sandbox. I installed HDP Sandbox successfully in oracle virtual box. But I am not able to get any desktop or eclipse or web browser on HDP sandbox. I tried by installing GNOME desktop but I was un-successful. I also tried to install VNC server and VNC viewer to get the desktop on the HDP Sandbox but it is not working.

Can any one help me on this issue please

1

There are 1 answers

0
Dennis Jaheruddin On

First of all, Cloudera and Hortonworks have merged. After the merger a new platform was released to succeed HDP and CDH. The new Cloudera Data Platform (CDP) is recommended to be used for anything moving forward.

Though there is not a sandbox as such, I found this excellent community article for getting a sandbox like experience with the trial: https://community.cloudera.com/t5/Community-Articles/How-to-create-a-Centos7-CDP-DC-Trial-VM-for-sandbox-learning/ta-p/300837

Now, for your main question: These kind of sandbox setups are a very tight fit, so I could imagine there have been some things done (like using all the resources or tuning the system in a way) that would make other things fail. However, you should not need to install anything like GNOME or Eclipse on the platform.

If you have something like Eclipse running on your laptop for instance, that is likely sufficient for most usecases. Keep in mind, this is just a sandbox, with an actual cluster you may want to fill in more of these things on client nodes.

For now, I would recommend you to really focus on the core. If you need to build something like spark jars, why not build them outside the platform and bring them in? That is also a common pattern for real clusters.

If you need to interact with the data with a GUI, why not connect to hive and such from the outside, again as one would typically do in a real scenario.


I realize some 'shortcuts' were available in the quickstart, but one should not need these so hopefully this will help you find ways to evaluate the platform in a sandbox like way without getting stuck on the difference with the old quickstart.