Capturing an image from Vendor X laptop and deploying to Vendor Y laptop?
We already use Fog for one of our customers, however they only utilize one type of laptop, and only windows. I was wondering, is it possible to capture an image from say, a teeny tiny dell laptop, and deploy that image to a much larger laptop from a different vendor? My concern is driver and hardware incompatibility. Basically, do I need to build a different image for each OS and each laptop variant we have? Or can I build the image off of my smallest drive available and deploy that to any computer? This significantly changes the amount of storage I need. Hope that makes sense, and thanks for any help.
Good to know. I’ve been rolling with the same script since Win7 and likely would not have looked at another way to do it short of it failing to run on a new version.
FOR /R “C:\Users%USERNAME%\Desktop\Drivers” %%G IN (*.inf) do pnputil -a “%%G”
Just as a point of note. If you have Win10 1703 and later you can use this command and not need the for loop. It does the same thing just in a single command line.
pnputil.exe /add-driver "C:\Drivers\*.inf" /subdirs /install
For way is needed before Win10 1703 and for Win7 the for-loop is the only way.
One way or the other isn’t right or wrong, its just different.
We do something similar to @george1421 but since we have a much more limited number of models to support (think 4 rather than 12) we inject the drivers before we capture the image from the VM.
We use enterprise Dell hardware so I download the driver CABs, then I extract all the drivers to a folder on the desktop called “Drivers”. I then run the following script to put them the driver store.
REM Will find and install all inf files in a folder called 'Drivers' on the current desktop FOR /R "C:\Users\%USERNAME%\Desktop\Drivers\" %%G IN (*.inf) do pnputil -a "%%G"
Keep in mind the are STAGED not Installed. So they are simply available to the machine at first boot to choose from.
I find this simpler than the injection process and the additional space used is negligible, though it doesn’t scale at large, but like I said we only have a handful of models. And now with Win10’s release cycle I can re-evaluate which drivers we put in the base each time we prep a new OS.
does this still apply with Linux and Mac
Linux is a bit of a different bird in that the linux kernels contain all of the needed drivers to boot on almost all common hardware. So for linux there is nothing really special you must do to prepare the golden image. Post deployment you will might need to set the computer name and some other install specific settings. But in general a linux distro using standard partitions (not LVM) is much simpler to deploy than windows.
I can’t speak for the Mac computers since I haven’t touch one since the mid 1990s. With some certain restrictions FOG does image those computers too.
Thank you, does this still apply with Linux and Mac?
Most Linux installs can be deployed to different hardware. Though there is no proof before you actually give it a try.
For Mac, how many vendors do you know?
@george1421 Thank you, does this still apply with Linux and Mac?
@alexnoel2 While my use case doesn’t exactly fit what you are doing, the concepts are similar.
We build our golden image on a VM so the image is hardware agnostic. Then we sysprep and capture the image with fog. We can take that single image and deploy it to our fleet of 12 different models of Dell computers. We do this by injecting the hardware specific drivers into the image just after FOG imaging is done and before the first reboot into Windows/OOBE. The drivers are copied over to the target computer using a FOG Post install script, and then loaded into the windows image by placing a command to call pnputil.exe in the setupcomplete.cmd file in windows. That pnputil.exe command load into windows the drivers left behind by the fog post install script.
This way as we add new models of computers we just need to install the dell drivers cab into the proper path on the FOG server.
Now hardware activation will be a different matter. This can be done in the setupcomplete.cmd file as long as the computer had windows 10 (same version) on it originally.
From a licensing standpoint you can not use an OEM sorce media to create a master image and redeploy it to XX number of computers, you will need a volume license media and key to do that and stay within the Microsoft EULA requirements.
cyannella last edited by
Only thing I can see being an issue is making sure you have the DriverStore for the larger laptop
saved off of it before you put the tiny image on it. Device manager will go nuts when it can’t find the updated drivers for the larger laptop. Do not blow them away before you image or you will have to waste time download them again.
Windows activation after resolving all the driver issues will be the next.
Make sure you expand the Large laptop hard drive to its max size.
Once you have a good working Large, capture it as you Large master image.