Which OS for SQL Server 2017 Standard? (new data warehouse)

  • My company will be setting up a new data warehouse server using SQL Server 2017 Standard. We're a mixed shop, running both Windows and Ubuntu, seems we could choose either. What would you recommend? Pros/Cons

  • Another detail, this'll be two servers in an Availability Group actually.

  • I would say Windows as the SQL AG story is more mature. Things should work fine on Linux, but I tend to avoid v1 versions of things like HA.

    However, feel free to test and see. I think Joey D'Antoni and Allan Hirt have been happy with the HA story, but I know it has been a journey for them.

  • Rekonn - Thursday, December 7, 2017 8:43 AM

    My company will be setting up a new data warehouse server using SQL Server 2017 Standard. We're a mixed shop, running both Windows and Ubuntu, seems we could choose either. What would you recommend? Pros/Cons

    When given such options, take both. 😉

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • What components of SQL will be being installed on this new setup?

    Do you need SSAS, SSIS and SSRS?  If so use Windows as they are not supported on Linux just yet, if its just the engine, look at what is not supported yet on Linux and see if you'll be using any of the features, if you are then you have to use Windows, if you're not then you can choose which one you'll be most comfortable supporting.

    https://docs.microsoft.com/en-us/sql/linux/sql-server-linux-release-notes#Unsupported

  • Considering that the OP stated that this'll be in a High Availablity Group, it might be worth noting there is a known issue with it on Linux. Not sure if it'll affect them, however:

    Certain algorithms (cipher suites) for Transport Layer Security (TLS) do not work properly with SQL Server on Linux. This results in connection failures when attempting to connect to SQL Server, as well as problems establishing connections between replicas in high availability groups.

    Details of a "resolution" can be found on the same page that Anthony linked to.

    Thom~

    Excuse my typos and sometimes awful grammar. My fingers work faster than my brain does.
    Larnu.uk

  • Steve Jones - SSC Editor - Thursday, December 7, 2017 11:25 AM

    I would say Windows as the SQL AG story is more mature. Things should work fine on Linux, but I tend to avoid v1 versions of things like HA.

    However, feel free to test and see. I think Joey D'Antoni and Allan Hirt have been happy with the HA story, but I know it has been a journey for them.

    the thing is, the Linux support is still evolving here.
    Tried and trusted for sometime now on Windows, the beauty is, you will be able to migrate to a new AG on a Linux system when the product is mature enough

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • Thanks everyone! Sticking with Windows. Didn't know about the linux HA issue, and that it's engine only for now (we'll need SSIS).

  • Rekonn - Wednesday, December 13, 2017 10:19 AM

    Thanks everyone! Sticking with Windows. Didn't know about the linux HA issue, and that it's engine only for now (we'll need SSIS).

    Integration Services is available together with Full-Text Search Features on Linux / vNext.  Although I wasn't aware of the issues with HAGs as Thom mentioned I also haven't experienced any problems in my test environment.

    For me the most compelling argument comes from Perry with regarding to support that is still evolving, so to that end if you aren't comfortable supporting under Linux, stay with Windows.

    Quick Edit:  I tend to separate BI components from the engine anyway so it may be the case that even though some features are not yet supported by Linux that there is no reason to use the engine on Linux and BI on Windows.

Viewing 9 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic. Login to reply