none
What is the advantage of using Vanilla Operating System in SCCM OSD versus if your create your image and "sysprepped" it RRS feed

  • Question

  • Hi,

    What are the advantages of using Vanilla Operating System in SCCM OSD versus if I create an image and modify it to how I want it to be, sysprep and make that our "golden image" to distribute to staff? I get the impression that this is the "ideal practice".  But why is that?  I know that space is one, duplicate Machine ID in Office is another, but what else?



    • Edited by DoBongSoon Monday, April 16, 2018 10:31 PM
    Monday, April 16, 2018 10:31 PM

Answers

  • Well if you follow the guide in the link I posted, that'll be your best way to make an image which is considered a "hybrid". It's just Windows only, but it also doesn't include everything (a thick image). Having an image which only includes Microsoft products such as the Visual C++ and maybe Office is easy to maintain and is pretty clean. You might notice in the example picture there's a suspend task. It's important that you build your reference image on a VM. Having the suspend task allows you to then create a checkpoint. The suspend should occur after all your automated tasks. That way in a couple months when you want to include patches, you apply the checkpoint, boot up the VM, run windows update and after all patches have been applied you shut down the VM and create a new checkpoint. Boot it back up and resume the task sequence. You'll end up with an updated image and very little work went into making that update.

    Just FYI the image will have sysprep run on it before it gets captured. Your best practice is to have an "administrative" deployment share. It's one that is configured purely for building images. Then you can more fully automate it and have it set to capture your reference image at the end of the task sequence, there's no need to make a separate capture task sequence.


    Daniel Vega

    • Marked as answer by DoBongSoon Wednesday, April 18, 2018 5:13 PM
    Wednesday, April 18, 2018 1:36 PM

All replies

  • The advantage of building a reference image is similar to that advantage of a cookie cutter versus having to hand form each cookie and expect them to be identical. The cookie cutter method is faster and more reliable.

    Here's a good place to start - Building a Windows 10 v1709 reference image using MDT


    Daniel Vega

    Tuesday, April 17, 2018 7:46 PM
  • Hi Daniel,

    Thanks! I will clarify my question more.  I understand that there are two ways of making an image

    1)To use a Vanilla Operating System -> Attach it to a task sequence ->  Attach all the programs in the Task Sequence -> Attach all preferences and customizations -> Create an image out of it then sysprep it. 

    2)Have an image with clean OS one time -> install all your standard apps, set preferences and apply cumulative changes to the OS and use this image moving forward.

    I prefer #1 because it us much cleaner and that's what I believe to be the ideal way of making an image.  On the other hand, my boss believes that #2 is better and easier because you already have the "golden image" and might as well pass that along and add the changes to that image moving forward. I can't explain well the importance of having a clean image... but maybe someone could say it better why #1 method is better than #2.

    Tuesday, April 17, 2018 10:16 PM
  • Well if you follow the guide in the link I posted, that'll be your best way to make an image which is considered a "hybrid". It's just Windows only, but it also doesn't include everything (a thick image). Having an image which only includes Microsoft products such as the Visual C++ and maybe Office is easy to maintain and is pretty clean. You might notice in the example picture there's a suspend task. It's important that you build your reference image on a VM. Having the suspend task allows you to then create a checkpoint. The suspend should occur after all your automated tasks. That way in a couple months when you want to include patches, you apply the checkpoint, boot up the VM, run windows update and after all patches have been applied you shut down the VM and create a new checkpoint. Boot it back up and resume the task sequence. You'll end up with an updated image and very little work went into making that update.

    Just FYI the image will have sysprep run on it before it gets captured. Your best practice is to have an "administrative" deployment share. It's one that is configured purely for building images. Then you can more fully automate it and have it set to capture your reference image at the end of the task sequence, there's no need to make a separate capture task sequence.


    Daniel Vega

    • Marked as answer by DoBongSoon Wednesday, April 18, 2018 5:13 PM
    Wednesday, April 18, 2018 1:36 PM
  • Thanks! Just FYI .. One problem I discovered with installing Office on a referenced image is that, the moment you open (Word for example), it creates a CMID, and this machine ID gets copied to other machines causing the KMS server to not increase the count.  Beside that, thanks for your information!
    Wednesday, April 18, 2018 5:13 PM
  • Don't open Office apps during the building of your image, also MDT will rearm Office when it goes through the sysprep/capturing process.

    Daniel Vega

    Wednesday, April 18, 2018 8:18 PM