Solution Repository - RDBMS or File-Based or Both?
In our shop, we are trying to setup multiple instances of Pentaho in our development environment so we can test out new configurations and/or reports without affecting other developers who may be working in parallel. We have these development instances connecting to a single database instance that contains the repository. We want to use the database-based repository so we can take advantage of the authorization built into Pentaho.
It was my understanding from the Pentaho documentation that the database or the file system would be the sole source of the pentaho solutions, not both, but from what we are seeing it seems to be a hybrid. Meaning, if your Pentaho instance doesn't have a file in the file system then when it "refreshes" the repository (for instance, when you create a new report), then the reports that you don't have locally get wiped out.
Is there a way to configure Pentaho to have the database be the master, or are we going to need to make sure that every instance of Pentaho has a mirrored copy of the solution repository? It appears as though the database stores the file in a CLOB. If it's a hybrid, what's the point of storing the file in the database if it essentially has to be on the file system as well?
Yes; the way it is designed to work, is that the file copy of the solution repository in the rdbms must be the same copy - i.e. mirrored, for every app server instance.
What's the point of storing the file in the database if it essentially has to be on the file system as well?
Good question, not really sure; However the database holds a lot more stuff than just the files in the filesystem.
Tags for this Thread