My last blog post on CAF primarily introduced our readers to the Oracle Content Accelerator Framework and its installation.  As part of my deeper dive I find that one key reference to using CAF properly is understanding the requirements for when a web catalog and/or RPD must be online or offline when cloning and synchronizing.  Therefore, as one of the first posts for the CAF deeper dive I am providing the matrix below as visual guide for what status a source or target web catalog and RPD must have  in order for CAF to work its magic.

Basically,  if the clone function is used, the RPDs for both the source and target must be offline.  The web catalog for the source can either be online or offline but the target web catalog must be online.   If you are using the synchronize function of CAF v1, then it doesn’t matter what the status of the source (selected) web catalog is as it can be online or offline.   However, since a the whole point of using synchronize is to compare an old RPD with a new RPD based on changes to logical columns and presentation layer columns and then modify the target RPD file, the target RPD must be Offline.  However, the original RPD (backup or early stage environment file) must also be offline due to some read/write functionality.

After my recent post on the Content Accelerator Framework and preparing for my deeper dive into the tool, I have begun looking at the inner workings of how to automate the tool for daily or weekly maintenance tasks.  This effort got me looking at some possible linkages between CAF and some existing OBIEE executables, nqXUDMLGen.exe and nqXUDMLExec.exe that reside in the ORACLEBI_HOME\Server\bin\ directory.

Andreas from Trivadis has done a spectacular job detailing the usage of XUDML in OBIEE, here.  I have also mirrored it here.

I should be posting my additional CAF findings this month.

Trivadis also has some other cool docs on BI which I have mirrored below:

• Hyperion Essbase/Planning Integration (German)
• Slowly Changing Dimensions (German)