There's nothing wrong with your design -- Reporting Services is built that way. Once your request is sent to the data warehouse, it's in the hands of that server. The process runs like this:
1. Data is requested from the database server (or other source) by Reporting Services.
2. The database server processes the request and generates results while Reporting Services awaits a response.
3. The database server sends results back to Reporting Services.
4. Reporting Services processes the results it received from the database server.
If Reporting Services gives up (either from timeout or user cancellation) before step 4, the database server continues its work for steps 2 and 3 anyway, sending the results back whether RS still wants them or not.
If user cancellations are a problem, you can check the ExecutionLog2 view to see who's aborting (look for TimeDataRetrieval > 0 and TimeRendering = 0). It may be one particularly impatient user.
↧