質問

Goal:
Apply a fresh deployed cube in the production environment.

Problem:
How should I enable to discover any problem in the cube when it is updated automatic once a week before it will be applied in the production phase?

Information:
In this context, there is no possibility to apply a user acceptance test and thest phase before production phase. There is only test and production phase.

役に立ちましたか?

解決

It sounds like you are having issues with data integrity/cleanliness. Users are human and will always enter bad data. Your job is to catch it before it causes an actual error. There are a few things you can do.

  1. Get BIDS Helper and run the dimension health check, dimension data type health check, and measure group health check to make sure your cube design is optimal. (Honestly, if you aren't using BIDS Helper for your cube, you are really missing out. I never build a cube without it these days).
  2. Update your ETL that creates the tables that are used by the cube to catch common errors (invalid data, missing surrogate keys on fact tables, etc.). You could write the errors to a file or table and have it email the issues to you or others.
  3. Set up another server with an identical database where the cube can process and then sync it with the production cube database so that a processing error due to bad data won't cause the cube to be unavailable.

I don't think there is a good way to test the cube processing without just doing it, but I'm confident you can find a way to take care of most of these issues before it ever hits the production database.

ライセンス: CC-BY-SA帰属
所属していません StackOverflow
scroll top