Every computer in my Windows 11 25H2 fleet is failing to install the September cumulative update. The August update installed normally, but the September release—and even the later out-of-band update—fails in the same way. The CBS logs mention a WLAN-AutoConfig manifest failure, although I'm concerned that may be a symptom rather than the actual cause. The installation reaches the reboot stage, progresses to a certain percentage, and then rolls back. I know error 0x800F0922 is often associated with an undersized EFI system partition, and I've already tried expanding that partition. I've also run DISM using installation media as a source, reset the Windows Update cache, and attempted an in-place upgrade. Has anyone encountered this failure pattern or found other checks worth trying short of reimaging the machines?
3 Answers
The operating system version is important here, so make sure all affected machines are actually on Windows 11 25H2 and have the same servicing baseline. Since you’ve already tried cache resets, DISM repairs, installation media, and an in-place upgrade, compare one failed machine with a clean 25H2 test installation. That can help separate a fleet-wide image or servicing-stack problem from an issue in the update itself.
Because the failure happens during the reboot and then rolls back, check the post-failure state of the system profiles. In particular, look for missing user.dat files under the system profile and default profile directories. Those missing files can sometimes point to a profile or servicing issue involved in the rollback.
There does seem to be a broader increase in failed Windows updates lately, and the September releases have been especially problematic. Since this is affecting the whole fleet and consistently rolls back after reboot, it may be more than a single damaged update package. I’d compare the rollback logs and setup results from a few representative machines rather than relying only on the WLAN-AutoConfig entry in CBS.log.
That’s what I’m trying to determine. I’m not sure whether Microsoft released a bad update or whether the fleet has a common servicing problem that this update exposed. I mainly want enough information to choose a recovery plan.
Yes, the update reboots, reaches a certain percentage, and then undoes the changes. Can you clarify exactly what the missing user.dat files would indicate and where you would check them?