views:

166

answers:

2

I have a SSAS 2005 database with 10 cubes. When I create the cubes from an XMLA script, it works fast, and I can browse cubes from SSMS, just as I should.

After processing the cubes during the night, eveyrthing starts to work VERY slow. Opening cube list in SSMS takes few minutes, and it is like this all the time. What can cause this?

+1  A: 

Check the usage of the server. A lot of times, this can occur because your cubes caused memory to spike and it hasn't released it. This would happen if your cubes had a lot of data processed over the night.

Try setting the memory limits of SSAS so that you reserve ~2GB RAM for the OS. If you can't spare 2GB for the OS, you need more RAM.

Eric
I checked the resources. When I try to expand cube list in SSMS the memsdrv process uses up to 10% of memory, and increases RAM by around 100 MB. But before doing it the SSAS is using around 80 MB of RAM. After a few minutes when the cube list is expanded the CPU drops to zero.Is there a way to set the minimal RAM amount that the SSAS should use?
Adam
+2  A: 

This should be a response to Adam's comment.

If you are running sql server 2005 sp2 or higher, you can try using the preallocate setting.

Make sure to read this whitepaper first so that you fully understand the ramifications of doing so: SSAS Memory Preallocation

JasonHorner