on 11-18-2008 2:46 PM
Hi experts,
I'm having some issues when trying to full optimize my application (on BPC 5.1). I get the following error:
Capacity must be greater than zero. Parameter name: Capacity
Save application and process from Analysis Services work without a complain but I'm not able to make it work from the application front-end.
The application is partitioned on a monthly basis, with 11 dims. I created a K2ServerDebug.log file on websrvr/bin for debug purposes but it seems that it's not creating the 4 other files
Any idea?
Thanks,
Ibai
What is the name of application? and what is the name of partitions?
It is very important to have name of partitions with this prefix Fact"NameofApplication".
Also make sure you don't have into any of fact tables values with Source <> 0. That can cause again issue for optimize.
Make sure that you have at least one record into one of fact tables.
Let me know if this help. If not please provide more details.
Regards
Sorin Radulescu
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Sorin,
Partitions follow Fact<CubeName>_<TimeSlice> convention. I'm able to run the full optimize when there are few records (~100 records) on the fact tables to process but it keeps failing if I do the same test with ~ 1,000,000 records.
Surprisingly if I run the optimization with compression it will complete without errors but I believe this way all partitions are processed regardless the data is new or not. I'm interested in measuring performance so I would like to complete this test withou the compression option checked.
I've run some queries on data to make sure that all the members exist and are not calculated and checked the source was equal to 0 .
I'm running this test on a (dev) monoserver environment (sql server 2005 + SSAS + BPC 5.1)
Regards,
Ibai
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.