Recommended way to create a master img?
goempie last edited by
What is the recommended way to create a hardware independed image?
I have read several option, one injects drivers through a script. Then i read things about wds (searched for a guide but don’t get what it does exactly and what the added value is)
Care to share how you guy’s make your hardware independed image?
boeleke last edited by
Hi there Goempie!
i’m on my last day in my internship.
And i must say… i get a superduper good tutorial what i followed.
Its only for W7 (can work for w10 – i don’t know).
I tried a couple of things for injecting the drivers.
One is putting them in the windows/inf – that works great after sysprep.
Or you can use Stand Alone Driver Utility (SAD3) !
I can speak for my company and how we do it.
We start out by using MDT to create a standard reference image in a virtual machine. In this MDT task sequence we only include the drivers necessary for the VM as well as the Dell WinPE deployment cab which contains the core drivers for our fleet of computers. We use MDT to fully automate the reference image creation. We have two task sequences. One is for a Windows thin image (Windows OS + windows updates), and the second is for a Windows Fat image (Windows OS + windows updates + standard software). Then at the end of the MDT task sequence we update the registry setting to tell windows to search for new drivers in the C:\Drivers directory and then in the C:\Windows\Inf directory. The complete reference image built takes about 5 hours. But its a bit like a Ronco build, you set it and then forget it. Come back in 5 hours and your reference image is fully built ready for sysprep.
Then we sysprep the image and power off the reference image
Then we pxe boot the reference image (running on the VM) and capture the image with FOG.
We test this newly captured image on all models we support. This is just a simple light up test (deploy, does it connect to the domain, can we log in, can we launch application x, y and z). Then we release the image for replication to all storage nodes.
We do this process once a quarter to keep the deployment image in FOG
You may notice that so far I have not mentioned drivers.
I have a post install script that determines the hardware model of the target computer and then copies the correct drivers from a directory on the FOG server to the C:\Drivers folder on the target computer. This coupled with the registry update tell the target computer where to find the drivers on the first boot post image deployment. If we add a new computer model to the fleet, we just download the Dell driver CAB and copy it to the drivers directory on the FOG server.
I create them in Virtual Machines, I have some batch files to automatically change certain things such as default user account picture and such.
It also adds ethernet/chipset drivers in a specific folder in C:\Windows\Inf so we always have network connectivity on first boot, which is necessary to load Snappy Driver Installer over the network for the rest of the drivers by launching the auto install batch file from firstlogon.cmd
Wayne Workman last edited by
This is older, but probably still valid: https://wiki.fogproject.org/wiki/index.php?title=Auto_driver_Install
There is more on this in the forums though, if you search around using different search criteria for it I’m sure you will find a lot of materials.