Guvnor Clustering

Clustering:

The following instructions are provided by Michael Finger (from Overstock.com) on how to enable clustering on Guvnor. This would generally only be done for availability, not performance. (If you need to have the rules on a high availability to runtime systems, then there are also other options, such as a local cache for the knowledge agent, and/or a proxy server etc...)

 

Instructions:

 

Here's a quick summary of how to get clustering working in Gunvor. We
use tomcat/linux/oracle, so your mileage may vary:

1) Drop war file in webapp dir in tomcat. Let tomcat explode the war and
then shutdown the server and delete the war file (otherwise tomcat will
overwrite your changes next time you start)
2) go to apache-tomcat-6.0.16/webapps/drools-guvnor/WEB-INF dir and edit
the components.xml file
Uncommit the following section so the file  looks like the following:

...................
 <component name="repositoryConfiguration">
         <!-- this is the directory I want my repo to be in -->
          <property name="homeDirectory">/opt/foo</property>
     .....
    </component>
.........................

3) In the foo directory refrenced above,  drop in a repository.xml file.
I grapped a default repository.xml file  out of the jackrabbit core jar
in the apache-tomcat-6.0.16/webapps/drools-guvnor/WEB-INF/lib dir. I
attached a default one.

4) Set up clustering in jackrabbit (Some resources I used:
http://wiki.apache.org/jackrabbit/Clustering  ,
http://jackrabbit.apache.org/jackrabbit-configuration.html#JackrabbitConfiguration-Filesystemconfiguration)