Hi,



I have a very small forecast cube (<10Mo) and some queries that are very
slow when executed over the cube towards http.



These queries are all build like this template:

With member

[member1] as '..'

[member2] as '..'

[membern] as '..'

Select { member1,, membern } on columns,

... On rows

from cube where .



One example is a query that return 365 days on rows, and fifteen calculated
members on columns. It takes :

10 sec via tcp/ip,

and more than 5 minutes on the same box but over http.



Using Execution location=3 does not solve anything.



Any tips?

Thanks,

Lionel

Hi Lionel,

Tips :-)
-Install Service Pack 3.
-Use the Execution Location setting of 3 and the Default Isolation Mode
setting of 1 TOGETHER.
-Set an appropriate value for the Large Level Threshold property.
-Group measures by usage.

See also "Optimizing Clients for Slow Network Connections" from:
http://www.microsoft.com/technet/tre...hNet/ProdTechn
ol/SQL/Maintain/Optimize/ANSvcsPG.asp

Also check this article there you have 15 calculated members this could be
interesting to test for you. Let me know if you need the fix (will be part
of sp4).

FIX: Analysis Client Sends Calculated Member Definitions When a Query Is
WGID:230
ID: 821479.KB.EN-US
http://support.microsoft.com/default.aspx?kbid=821479

HTH,
Bas


"This posting is provided "AS IS" with no warranties, and confers no
rights."