Confused about all the files

[dean12]

I migrated AnalysisServices to use a SQL database. I named that database AnalysisServices. I also have a database that is the data warehouse and it contains all the raw data - it's name is DataWarehouse.

Now I went to ARCHIVE in AnalysisServices Manager and when it does it's thing, it's processing though 100+ files with names like ....rigid, ...fact.map, .....pdr, etc.

The ARCHIVE created a .CAB file and it says everything worked just fine. Question is:

1. What are all these 100+ files?

2. If I want to migrate this to another server, then do I need the .CAB file, the AnalysisServices database, and the DataWarehouse database? I'm confident that I need the DataWarehouse database as it has scripts and such that are needed to put the raw data together. I'm also thinking that I DO NOT need the AnalysisServices database as that might be on the destination server already with perhaps other analysis services information about other cubes, etc.

Can anyone clarify? Appreciate it.

 

[MDXer]

The files you are seeing are the MOLAP Data structure files These contain the processed data. When you archive an OLAP database in AS is cabs all the files and metadata for that cube into a single archive. This is all you need to move or restore AS databases between servers. Migrating the Repository to SQL Server has no impact on how or where the processed data is stored, it only determines where the Metadata is stored. When you restore a OLAP DB to a server the metadata is then written to that servers repository.

The data warehouse is the relational Structure that the cube is built from. Typically you don't move your warehouse from server to server.

Analysis Service Cubes if created in a MOLAP structure are 100% self contained unless drill through is enabled. This means that without drillthrough the cube will function without the support of the DW until it is time to reprocess the cube. Drill through does require the supporting warehouse however since you drill from the cube into the relational system to bring back the records the aggregate value represent. MOLAP is becomming and is my preffered storage for this reason and the speed factor a well designed molap cube is extremely fast often sub-second.

This means that if OLAP ServerA and OLAP ServerB can both access the Datawarehouse then you do not need to move your warehouse.

HTH


Shoot Me! Shoot Me NOW!!!
- Daffy Duck