This Demo is used to deploy automatically Hadoop with BlobSeer on Grid5000 under the deployed environmentand, and run various Hadoop applications and benchmarks tests. Here is the instruction for the installation and use of the Demo.
The guide is divided into different pages. The first three pages are mandatory to be read. As you move on, in sequence, you must complete each page's instructions. By the end of each page you will have been able to test your work using given examples such that you know you can progress.
The other pages are optional. However, it is recommended to also have a look at them.
The user should have at least one Grid5000 account to make the Demo work. This instruction lists all necessary SSH configurations to connect to Grid5000 and use the Demo.
On this page you will see how to reserve nodes, then to boot our customised image with HDFS/BSFS and finnaly allow discussion between nodes.
After the previous section done, node are reserved, custom image is deployed on them and they can communicate each other. So the user should be able to run some tests to investigate the performance of BSFS. This page includes the explanation of provided test setups, and some examples of running the tests. We provide two kinds of test to compare the performance of BSFS and HDFS including application tests and benchmark tests.
We also provide some mechanisms to visualize the results of chained tests. On this page, we give the instruction of the visualization tool.
In case of failure, the user can follow the instructions on this page.
With different aim, the Demo should be run in different ways. This page gives tips of using the Demo in various use cases.
Of course, the user can customise the deployment of BlobSeer and Hadoop-BSFS after Grid5000 nodes are launched. This instruction shows how to customise the deployment on a Grid5000 node instead of Grid5000 front-end.
Any idea about improving this Demo are welcome. If the user wants to get his/her hands dirty, here is the instruction to create a Demo image from scratch. Based on this instruction, the user may further optimize the Demo.
Here is explain how to reserve, deploy and modificate an image.
Here are given benchmark used by IBM that we can run on grid5000