If you are unable to create a new account, please email bspsoftware@techdata.com

Author Topic: Dynamic cube performance really slow  (Read 3783 times)

Offline zgeorge

  • Full Member
  • ***
  • Join Date: Feb 2018
  • Posts: 25
  • Forum Citizenship: +0/-0
Dynamic cube performance really slow
« on: 18 Apr 2019 07:33:23 am »
Hello Cognos Dynamic Cube masters!

We have created dimensions in our dynamic cube designer that are built to allow for drill downs in specific reports. We have noticed some really slow performance when using these dimensions and have come across the error messages in the image attached. Can anyone provide some insight on what I need to do to fix these performance issues based on the error messages?

Offline bus_pass_man

  • Statesman
  • ******
  • Join Date: May 2008
  • Posts: 520
  • Forum Citizenship: +47/-0
Re: Dynamic cube performance really slow
« Reply #1 on: 19 Apr 2019 02:38:01 pm »
I think you'll find that those are warnings, not errors. 

The messages are trying to identify situations which could be problematic but are not deterministically so.

Without knowing your data warehouse, the validity of the design of your dynamic cubes (such as the design of the problematic dimensions and your aggregates setup), the performance problems you're having, the scenarios in which the problems, and mundane things such as what sort of hardware you're trying to use, all I can say is download the dynamic cubes redbook and, possibly, the dynamic query redbook and work your way through the troubleshooting sections and also to read the text of the messages and try to understand whether they are valid or the design trade-offs that you need to accept.

'Fact acct sls drill' sound like an odd name for a dimension, especially the fact part.  The fact that you have a level in it  called internal distribution channel, another level in it called product type, and another called plan core business unit suggests that something is particularly odd with that dimension.  Those sound like levels from completely different dimensions rather than levels of a single dimension.