Categories

Versions

You are viewing the RapidMiner Hub documentation for version 10.0 - Check here for latest version

Installation

The Scoring Agent is distributed as docker image. It uses the SCORING_AGENT_HOME_DIR environment variable, which is used as the home directory for the Scoring Agent and resides in a persistent docker volume (default location in the docker image is /scoring-agent/home) -- let's call the new location $SA_HOME. Before you start the Scoring Agent, you will need to provide a valid license.

Provide a license

To install a Scoring Agent license visit my.rapidminer.com and copy the license key to your clipboard. Afterwards navigate to $SA_HOME/home/resources/licenses/rapidminer-scoring-agent and create a new license file with the name rapidminer-scoring.lic. Paste the content of your clipboard into the file.

Besides that, you could alternatively add the LICENSE_LICENSE environment variable with the contents of your clipboard and the RTS will pick the license up from its environment.

Configure Scoring Agent

To configure the behavior of the Scoring Agent, you can set different environment variables. See also the table of system settings.

The default port of the Scoring Agent is 8090. If you want to change the port modify the SERVER.PORT environment variable and restart the Agent.

Install extensions

To install RapidMiner extensions go to $SA_HOME/home/resources/extensions/ and place the desired extension JARs into the folder.

Configure extensions

To configure extension properties open the $SA_HOME/home/config/rapidminer/rapidminer.properties file and change its contents. The Python Scripting extension, for example, can be configured by uncommenting the line rapidminer.python_scripting.path and pointing it to the binary of the Python installation:

 ##
 # Set the properties of RapidMiner and available RapidMiner extensions.
 # Below are two examples on how to configure the Python and R scripting extension.
 ##

 ##
 # Set the path of the python binary for the Python Scripting extension.
 #
 rapidminer.python_scripting.path=/usr/bin/python

 ##
 # Set the path of the R binary for the R Scripting extension.
 #
 #rapidminer.r_scripting.path=/path/to/r_binary

Start the Scoring Agent

To start the Scoring Agent, spin up the underlying docker image. Once the Scoring Agent has successfully started, you will see the following log lines:

INFO 6221 --- [main] s.b.c.e.t.TomcatEmbeddedServletContainer : Tomcat started on port(s): 8080 (http)
INFO 6221 --- [main] c.r.execution.scoring.Application     : Started Application in 20.226 seconds (JVM running for 22.243)

What's next?

Next you can create a new deployment from RapidMiner Server or install an already existing deployment.