Installing WebSphere® Portal can be a time consuming process. There is a great deal of software involved and even though the actual installation process is largely automated, the elapsed time required to simply lay it all down can be significant. For availability and throughput reasons, it is also common to go beyond a single instance of WebSphere Portal by clustering several of them together. However, the additional installations required to complete the cluster act as a multiplier, driving elapsed time quite high. Wouldn't it be more efficient to go through the installation process once, then clone that instance as many times as required across the cluster? We think so, and this paper reveals an approach for doing just that.

To put the cloning phase in context, we show the entire process from beginning to end. We start by walking through a couple of sample installations—first a default (recommended) install and then a custom install for those who need to follow that path. Then we review the cloning process, including some XML and bash shell scripts to automate much of the busy work. Next we cluster the WebSphere Portal images together using WebSphere Application Server Network Deployment. Finally, we migrate the datasource from CloudScape to DB2® on either Linux or z/OS® . The end result is a fully functional and serviceable WebSphere Portal cluster deployment on Linux for System z™ .

WebSphere Portal Server Cloning on Linux for zSeries and z9

Introduction

Overview of WebSphere Portal

Installing the first WebSphere Portal instance

Custom installation

Cloning

Copy the WebSphere Portal image

Modify the WebSphere Portal configuration

Appendix A

Appendix B

The team that wrote this Redpaper