4. Eucalyptus and Twister on FutureGrid¶

4.1. I. Introduction¶

This tutorial will show you how to use Twister under Eucalyptus on India, FutureGrid.

4.2. II. Prerequisite¶

Follow tutorial Using Eucalyptus on FutureGrid to learn how to install and use the Eucalyptus client tool to access resources on India, FutureGrid.

4.3. III. Download FutureGrid Eucalyptus Twister Tool¶

This tool is a set of python scripts. They can provide a pre-configured Twister environment, and also can terminate the environment. Pleasecdownload the tool in the attachment below.

4.4. IV. Start Twister Environment¶

To start a Twister environment, execute the following program:?

rteleft

$ python fg_euca_start_twister.py [-k user key] [-i public key file path] [-n number of instances][-t instance type]

e.g.
$ python fg_euca_start_twister.py -k userkey -i userkey.pem -n 3 -t c1.medium

Here,

• -k is the user key name generated by the euca-add-keypair step in the Eucalyptus tutorial.
• -i?is the private key .pem file path. It is also?generated in the euca-add-keypair?step in the Eucalyptus tutorial.
• -n is the number of instances for starting.
• -t?is the type of image.

The following is an execution example:

Once the script is executed, the user can get a prepared Twister environment. Then, the user can follow the instructions provided by fg_euca_start_twister.py?to start ActiveMQ on the assigned node, and also start the Twister environment (could be on any node just applied).

4.5. V. Terminate Twister Environment¶

To terminate a Twister environment, execute the following command:

 rteleft

$ python fg_euca_terminate_twister.py

4.6. VI. Run Twister-Kmeans¶

$ cd /opt/Twister/samples/kmeans
$ ant
$ cd ../../lib
$ mv Twister-Kmeans-0.9.jar ../apps/
$ cd ../bin/
$ chmod a+x twister.sh
$ ./twister.sh cpj ../apps/Twister-Kmeans-0.9.jar

$ cd /opt/apache-activemq-5.4.2/bin/
$ activemq console

$ cd /opt/Twister/bin
$ ./start_twister.sh

$ cd /opt/Twister/bin
$ ./twishter.sh mkdir kmeans

$ cd /opt/Twister/samples/kmeans/bin/
$./gen_data.sh init_clusters.txt 2 3 /kmeans km_data 3 30000

$ ./create_partition_file.sh kmeans km ../samples/kmeans/bin/p.pf

$ ./run_kmeans.sh init_clusters.txt 3 p.pf