locked
Push changes to another cube RRS feed

 > 

  • Question

  • Question
    Sign in to vote
    0
    Sign in to vote

    Hi everyone.

    I have a SSAS cube which is connected to Power BI. To update data in cube, I planned to process it several times as scheduled procedure. I came up with a procedure to have two cubes, one connected to Power BI and another to process and update data. As I understand from it, I need to have two SSAS servers to connect between them. Is it correct? Is it possible to have such option in one SSAS server?

    Also, I'm wondering about synchronization of cubes, is it possible to have it scheduled and only sync of origin cube is processed?

    Regards,

    Yerkhan

    Tuesday, August 14, 2018 7:29 AM

Answers

  • Question
    Sign in to vote
    0
    Sign in to vote

    Hi Yerkhan,

    Thanks for your question.

    >>>I have a SSAS cube which is connected to Power BI. To update data in cube, I planned to process it several times as scheduled procedure. I came up with a procedure to have two cubes, one connected to Power BI and another to process and update data. As I understand from it, I need to have two SSAS servers to connect between them. Is it correct?
    No, it is not correct, you can create two cubes in one SSAS server, it is not necessary to install two SSAS instances.

    >>>Is it possible to have such option in one SSAS server? Also, I'm wondering about synchronization of cubes, is it possible to have it scheduled and only sync of origin cube is processed?
    Yes, It is possible, please refer to below blog talking about this issue:
    http://richardlees.blogspot.com/2009/05/new-analysis-services-database-version.html

    Best Regards
    Willson Yuan
    MSDN Community Support
    Please remember to click "Mark as Answer" the responses that resolved your issue, and to click "Unmark as Answer" if not. This can be beneficial to other community members reading this thread. If you have any compliments or complaints to MSDN Support, feel free to contact MSDNFSF@microsoft.com

    Tuesday, August 14, 2018 7:56 AM

All replies

  • Question
    Sign in to vote
    0
    Sign in to vote

    Hi Yerkhan,

    Thanks for your question.

    >>>I have a SSAS cube which is connected to Power BI. To update data in cube, I planned to process it several times as scheduled procedure. I came up with a procedure to have two cubes, one connected to Power BI and another to process and update data. As I understand from it, I need to have two SSAS servers to connect between them. Is it correct?
    No, it is not correct, you can create two cubes in one SSAS server, it is not necessary to install two SSAS instances.

    >>>Is it possible to have such option in one SSAS server? Also, I'm wondering about synchronization of cubes, is it possible to have it scheduled and only sync of origin cube is processed?
    Yes, It is possible, please refer to below blog talking about this issue:
    http://richardlees.blogspot.com/2009/05/new-analysis-services-database-version.html

    Best Regards
    Willson Yuan
    MSDN Community Support
    Please remember to click "Mark as Answer" the responses that resolved your issue, and to click "Unmark as Answer" if not. This can be beneficial to other community members reading this thread. If you have any compliments or complaints to MSDN Support, feel free to contact MSDNFSF@microsoft.com

    Tuesday, August 14, 2018 7:56 AM
  • Question
    Sign in to vote
    0
    Sign in to vote

    Hi Yerkhan,

    Willson has already provided you the answers. As an alternative you could use the logic of swapping cube databases. At a high level, the approach will be as mentioned below

    1. Deploy 2 cube databases SSAS and SSAS_Process to AS instance.

    2. Using a script component task in SSIS, process SSAS_Process database.

    3. Add another script component task which will have the logic to swap databases i.e SSAS_Process once refreshed will be renamed to SSAS and existing SSAS database will be renamed to SSAS_Process database.

    HTH,

    Ram

    Please vote as helpful or mark as answer, if it helps

    Tuesday, August 14, 2018 9:00 AM
  • Question
    Sign in to vote
    0
    Sign in to vote

     I came up with a procedure to have two cubes, one connected to Power BI and another to process and update data. 

    Why did you come up with this process? What problem are you trying to solve?

    Cubes are always processed in transactions. So while the cube is being processed the users will only see the previous set of data and only once the transactions commits will they see the new set of data. Never at any point is it possible for users to see partially processed data (as long as your processing is all done in a single transaction)

    http://darren.gosbell.com - please mark correct answers

    Wednesday, August 15, 2018 3:16 AM