FOG cloning: stick with Sysprep or use a SID changer?

November 22, 2026 4 answers
Asked by Zavdelo.7r

We use FOG Project to reimage university workstation labs. Our Windows 11 image is built in Audit Mode and generalized before capture, but Sysprep keeps failing on built-in Appx packages.

Would a post-clone SID changer be a reasonable alternative? I would also like to hear how people keep deployments manageable without a paid remote management platform.

Answers

Vordelo-3m

I would keep Sysprep /generalize /oobe before capturing a reusable Windows image. A SID changer is not a substitute for the supported generalization process, and duplicate local machine SIDs are not the whole cloning problem anyway.

Watch for other machine-specific state too, especially domain membership, certificates and management or security agent identities. Check each vendor's cloning guidance rather than assuming Sysprep resets every third-party agent.

Reply to this answer
n.651!vexlano

For the Appx failures, start with the Sysprep logs and identify the exact package and user involved. A package installed for an existing user can get out of sync with the provisioned package intended for new users. Removing only one side may not fix it.

Keeping Store apps from updating during image preparation can help, but an offline build is not a guarantee that every package state will be clean. Test any package removal on a disposable image and make sure you are not removing an app your users need.

Zavdelo.7r:

The build is already offline, so checking the package state for existing users as well as provisioning sounds like a useful next step.

Reply to this answer
4m~cobblevisor_726

Consider a much thinner image. Deploy a clean Windows base, then use FOG snapins, GPO scripts or scripts from a local share to install applications and apply configuration. That lets you update installers without rebuilding the whole image whenever an app changes.

For lab reimages, test a full batch rather than one machine. Local installers may reduce repeated internet downloads, but the deployment share, network and older disks can still become bottlenecks.

Nexdori-5k:

A thin image still needs a tested process for drivers, updates and application dependencies. Moving installs after deployment changes the work rather than making it disappear.

Reply to this answer
_b71~quedrano.5

You could also look at a fresh Windows installation workflow using WinPE, or a tool such as OSDCloud, instead of capturing a customized image. That avoids maintaining a large golden image, though the driver coverage and download behavior need testing on the older models before relying on it for a whole lab.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions