Plan for the Portal Personalization Server

 

+

Search Tips   |   Advanced Search

 

You might want to use the Personalization rules engine in an environment outside of IBM WebSphere Portal. For example, you could use Personalization in a Web service which provides a personalized XML content stream. Alternatively, you could use it to generate an RSS feed. In these cases, Personalization is used as a service which is consumed by other software components on other servers. A full WebSphere Portal installation incurs some administrative and infrastructure overhead. When portlet rendering is not required, such as when the output of a service is an XML or RSS feed, you could install the Personalization rules engine in an environment outside of WebSphere Portal. You configure a Personalization standalone server - an Application Server with the Personalization rules engine (without WebSphere Portal ) - using the installer on theWebSphere Portal V6.0 CD # 5.

The servlet which receives the publishing request can also run outside of WebSphere Portal. With the exception of pznwpsruntime.jar, which includes the WebSphere Portal user resource collection, all the Personalization run-time libraries can run outside of WebSphere Portal.

We can install the repository, which stores the rules and campaigns, outside of WebSphere Portal. In this configuration, the DB2 Content Manager Runtime Edition repository does not use WebSphere Portal Access Control.

The only components which are not installed when running outside of a Portal system are the Personalization portlets, including the authoring portlets and the Personalized List portlet. As a consequence, once artifacts are published to a Personalization rules engine without WebSphere Portal installed there is no way to view those artifacts directly on that system. You need a WebSphere Portal installation, either a test environment installation or a full WebSphere Portal installation, in order to author rules and campaigns.

Publishing to a server which does not have WebSphere Portal installed is no different than publishing to a server where Portal is installed.

 

Disable Text Search

The Personalization Server has the same requirements as WebSphere Portal in regards to the WebSphere Application Server product.

The following warning messages appear in the standard out log when using Personalization without Portal installed and in the Rational Application Developer test environment.

ConnectionMan W J2CA0075W: An active transaction should be present while processing method allocateMCWrapper.
ConnectionMan W J2CA0075W: An active transaction should be present while processing method initializeForUOW.

To work around this problem, disable text search from IBM Content Manager by editing...

wps/jcr/lib/com/ibm/icm/icm.properties

...changing...

jcr.textsearch.enabled=true

...to...

jcr.textsearch.enabled=false

Alternatively, if you require text search for the installation, we can hide this message by changing the j2c.properties file in was_root/properties as follows:

  1. Uncomment the following lines:
    <cm-properties>
        <manageCachedHandles>false</manageCachedHandles>
        <logMissingTranContext>true</logMissingTranContext>
    </cm-properties>
    

  2. Change true to false for line:

    <logMissingTranContext>false</logMissingTranContext>

After installing, restart the server.

 

Database considerations

Personalization no longer has an independent administrative or authoring database. Personalization uses the DB2 Content Manager Runtime Edition for storage of rules, campaigns, and other objects both during the authoring phase and runtime phase.

Two optional functions of Personalization require the use of a database.

  • The Feedback function uses a database to store logged information. The amount of database space that is required for logging depends on the amount of traffic to the site. The amount of data that is logged per login-enabled page can vary.

  • The LikeMinds recommendation engine uses a database to store information gathered through the logging APIs. LikeMinds uses this information in order to make recommendations.

 

Related information