I have a firecuda m.2 that I use for the pc installing windows, I also have a samsung m.2 to store all my files and folders, there is no way in hell am risking harm to my samsung m.2 for this build or any other build, this is typical of microsoft this, you would really think they would have no problems with the higher builds sometimes i just wonder if they do this deliberately
It looks like a weird bug in the servicing of the 26040 and newer. The Servicing copies all the required files of the 26052 to the host in its temp folder including the correct ucrtbase.dll. But for some kind of reason it is not used and it tries to use the ucrtbase.sll from the host. And that leads to the crash of the dismhost. Don't know how to fix this, especially since the problem isn't related to the converter script. This also happens in the terminal if you try it manually. Some of the DISM commands simply won't work if the host isn't also 26052. And the problem with the LCU and the Core/CoreN to 26052.1100 is also strange. Both images are inconsistent from the start. Looks like both .esd files have a bug. Withe the Pro/ProN there ist no problem. Maybe ist can be fixed, but I don't know where in the image of the Core/CoreN it is determined which edition is active. But perhaps this bug is resolved with the next Dev. I wait. Only testet with de-DE.
@TS2008 That doesn't explain why framework Apps can be integrated successfully even full Apps, for some users, first or second app succeed, the rest fail LCU fail for Core/CoreN probably because they include staged target editions, while Professional/N doesn't and since LCU include new version of the editions packages, the stack get confused and messed up, it flag new version as active before superseding rtm version, or so same for Server, Standard and Datacenter are upgradable, Turbine is not (maybe someone can test if LCU fail for it too) Win10 1903 stack had a bug when integrating LCU to WinPE, it tries to remove the RTM packages before installing the new version packages this can be used to enumerate editions state (run in command prompt) Code: SET "RegKey=HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Component Based Servicing\Packages" SET "Pattern=Microsoft-Windows-*Edition~31bf3856ad364e35" FOR /F "TOKENS=8 DELIMS=\" %A IN ('REG QUERY "%RegKey%" /f "%Pattern%" /k ^| FIND "~"') DO @REG QUERY "%RegKey%\%A" /v "CurrentState" can also be ran against offline mounted hive
This also gives me a headache. It's a bad bug. Depending on the app, maybe different RPC servers are used. I don't know. Ok. I now understand. Removing the staged Pro Edition solved the problem with the LCU. Now only one edition ist active. Thanks for pointing this out. Code: Index : 1 Name : Windows 11 Home Description : Windows 11 Home Size : 16.074.810.312 bytes WIM Bootable : No Architecture : x64 Hal : <undefined> Version : 10.0.26052 ServicePack Build : 1100 ServicePack Level : 0 Edition : Core Installation : Client ProductType : WinNT ProductSuite : Terminal Server System Root : WINDOWS Directories : 19644 Files : 92557 Created : 03.02.2024 - 14:30:25 Modified : 10.02.2024 - 22:20:08 Languages : de-DE (Default)
did you uninstall the driver completely. for my system its samsung nvme driver for sure. even after updating reinstalling the driver result in the same error on boot.
oops!.. didn't realise that I had also stuck the Samsung driver on the OS drive as well, believed I only applied them to the RAID (which at least, was the original intention) some (rather interesting) notes on my Win11 "Hardware Requirements Not Met" failure scenario... 1... DirectStorage requirement --> an NVMe SSD that uses the "Standard NVM Express Controller" <-- basically, Microsoft's own (lower performance) driver. ↑ Gaming feature... Not strictly a requirement and easily resolved by swapping out the Samsung driver with the Microsoft one. (One could make an assumption here, that doing the same could (or should) resolve the storport.sys issue we are apparently having... something for another day.) 2... CPU requirement --> Hardware support for Hypervisor-Enforced Code Integrity (HVCI)... AMD's Guest-Mode Execute Trap for NPT (GMET). ↑ This one I find rather comical as there is no difference (that I could find) between the 1950 (unsupported) and the 2950 (supported), other than the size of the Silicon DYE they're printed on.
I replaced all my 19044 DISM and various ucrt files with current ones, taking permissions away write permissions from the OS and making them read-only so they won't get "repaired." Works fine. Another thing I noticed was if you close out the convert process at any point DISM has started work and before it's done, good chances you'll get the RPC error. I would have one build work, then start another, cancel it, and RPC errors began. I don't know if it was cleaning mountpoints or the file replacement that did the trick. I was hopping drives until none would build a proper ISO anymore (would also get "there is already a mounted file" errors after closing failed converts and deleting their files), but I did both over several hours and then rebooted. Problem ultimately solved.
hello i wanna upgrade 22635.xxxx to this build without going into the insider channels.. and keep my data... what i have to do to get that ??
Try to download from uup then mount iso and try to upgrade if no go i would probably use rufus to and disable tpm checks then try to upgrade again. i dont think you will be able to downgrade if you wanted to without loosing your data so maybe do some backup just in case
the upgrade will work.. but cant keep my data without entering insider.. i know i could do an cleaninstall.. but i loose my data there