Probably been answered somewhere before, but this is one for the WDS experts out there.
We've got WDS deploying images fine and I'm familiar with WDS, WAIK, WSIM etc. and unattend files. We are currently using WDSClientUnattend.xml which is specified in the server properties to specify an administrator user to perform imaging, and to wipe and partition drives with a single partition before imaging. This is working fine, but now I want to specify different partitioning options for different machines. What I want to achieve is to have my staff laptops image onto a single partition that fills the entire harddrive, but for our classroom laptops I'd like to create a partition for the O/S and applications, and a second partition/logical drive to act as a 'scratch' drive.
I understand that I can't achieve this through WDSClientUnattend.xml at the server level, and would have to remove this, but can I do it at the image level instead?
Have you looked at MDT, that allows for the usage of existing images and WDS but also gives you many more choices at the clientside when imaging. We use a vaguely customised MDT share in all our places to dish out images so that reinstalling is just a carefully places F12 press away. I have not dealt that much with raw WDS, I'm sure its doable in there but I just find MDT is a much easier solution when you want to customise anything.
Could you just add the same image to wds with a different name and attach a sysprep file that specifies different drive formatting
Thanks Synack. I've played around with MDT but couldn't get my head around it. Maybe time to go back and look again. I think I'm right in saying that you use MDT to configure your images, but you'd still deploy these with WDS?
Yes, and we do already have multiple images in WDS but I'm confused about the point at which partitioning takes place. With a WDSClientUnattend.xml file specified in the WDS server properties, as soon as we have chosen an image to deploy, the drive is wiped and a single partition is created (as specified in the WDSClientUnattend.xml file), and this is fine if we want every machine that we image to have a single partition. But now we want 2 partitions on certain machines/images. So if I deploy image A it wipes the disk and creates a single partition, but if I deploy Image B it wipes the disk and creates 2 partitions. But obviously I can't achieve this with a single WDSClientUnattend.xml file specified at the server level. So I should:
Originally Posted by sted
1. Remove the WDSClientUnattend.xml from the server? (or I guess I could modify it to remove partitioning info)
2. Associate imageunattend.xml files on the individual images, with the WindowsPE\x86_Microsoft-Windows-Setup_neutral pass to specify partitioning? (I tried this and it didn't work, I was still prompted to manually create partitions)
Pretty much, MDT takes your images and injects the needed configuration stuff and drivers after applying the image. It is all kicked off by a custom WindowsPE image that MDT creates. This can be started off through WDS for network boot or even a CD/DVD/USB key, all of which you can generate from the app. This can either point to a network share or you can generate whole install sources to DVDs or external HDs for remote sites.
Originally Posted by earlyriser
I can give you a hand with MDT stuff if you'd like. My setups only really use the basic features but there is room to tie it all into a DB making for automaticlly configured reinstalls.
For mantainability I find MDT really good as it handles all the driver stuff and allows you to use a single generic image that can be pushed out in multiple ways to different hardware. The package addition is also quite handy for when you have a stack of the same kind of machines as you can just tack on all the laptop specifig software after the base image. Updating these for new installs is a snap to.
You can also set it to use multicast through WDS which can make it much faster for multiple stations if your network is up to it and drivers are not evil.