I have a line of communication open with someone in Microsoft’s Product Support team, but there are long pauses between response times. So, maybe a fellow doper can help me out on this.
I have a few ROLAP cubes, each containing about 5 or 6 dimensions. All of the corresponding dimensions are also using ROLAP for their storage mode. In an effort to achieve real-time ROLAP, I’ve configured these dimensions (and cubes) to use Proactive Caching.
Here is a simple scenario, which is resulting in multiple deadlocks. Say I configure two of the ROLAP dimensions to use Proactive Caching, with the Scheduled Polling notification scheme. I specify “select getdate()” as the polling query. (This is to ensure that the cache is cleared/rebuilt upon every execution of the query.)
Running Profiler on Analysis Services, I see multiple deadlocks occurring on these two dimensions. However, I do not see any deadlocks occurring on the relational DB side of things. The deadlocks are causing the dimensions (and the cubes to which they belong) to become completely unresponsive.
Has anyone else encountered a problem like this with their setup? Or can anyone offer suggestions on what might be causing the deadlocks? I’m at a complete loss.
LilShieste