October 26, 2009 at 11:13 am
I have a decent sized cube that sources it's data from a 2005 SQL Server DB. We process the cube on two different Analysis Services servers. One located where ther source DB is, the other on a separate server. Each cube processing takes about 40 minutes each day and are run one after the other. On some days however it will run for 200 minutes. This used to happen to both AS servers on the same day. We switched the order of the processing and now when one runs long the other doesn't. Sometimes it's the first one that runs long, sometimes the other. There is nothing else happening on the SQL Server DB at the time of cube processing. About 80% of the time everything works fine. Profiler traces don't reveal anything out of the ordinary on slow days. Any suggestions?
January 16, 2011 at 11:55 am
You have to tinker with the cube database properties until you get the correct settings for your circumstances and data. What I have found is that as the cube grows you need to reset these settings on a regular basis. There is a good document available if you type 'SSAS 2005 Best Practices Download' in your search engine. It tells you what each setting does. Incrementally change these by 1 or 2 and eventually you will get the optimum performance setting for your cube processing.
NB - always document the changes you have made.
July 27, 2011 at 4:48 am
hi
I have a similar situation One day it is taking only 30 min.But the other day it is taking more than 3 hrs .
How did you solved this actually?
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply