After upgrading our Team Foundation Server environment, we ended up with 2 collections each containing their own process templates, build server and so on…
The concept of a Team Project Collection is a really great improvement in the TFS solution, but there is one caveat: schema conflicts can occur when a set of attributes for reportable fields differs across team project collections.
When a schema conflict occurs, data that is associated with that schema cannot move into the data warehouse and the SQL Server Analysis Services data cube. You must correct all schema conflicts to unblock processing of the associated data for the warehouse and to enable the associated reports to display current data.
All reportable data from all team projects that are defined in all project collections for a deployment of Visual Studio Team Foundation Server is written to a single relational data warehouse. Data from that warehouse is then processed and written to the cube. Collecting data into a single data warehouse supports reporting across team project collections. However, because fields are managed distinctly for each project collection, schema conflicts can occur when different definitions are assigned to one or more attributes of a field that is assigned the same reporting reference name.
More info here: