I have several Windows 10 computers that run our App which controls our hardware. Each run of our App takes several days to complete. When a Windows Update comes in, it reboots the computer which kills the process and ruins the run. This causes us to have to throw away the incomplete results and start over, which costs over $1000 in consumables, several hours of effort and a delay of getting the results to our customer. I have tried setting the registry entry:
HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Microsoft\Windows\WindowsUpdate\AU\NoAutoUpdate to 1
Setting the Windows Update service to disabled
which is supposed to prevent updates, but something in Windows resets the Windows Update service from disabled to manual, which allows the update. I have since discovered a couple more registry entries which seem to be what I need, as well as other processes that purport to do what I want, but my question is:
How do I test them?
One way is to simply wait until the next update (which I did with the above attempt since I trusted it and had no way to test, but since it didn't work, another run, and $$, was wasted).
Is there a "Test Update" I can "push" to verify that it doesn't update my computers?
Another option is to add a script to the beginning of running the App to tell Windows to delay any updates for 7 days (which some scientists do manually before each run) but again, how do I test if that will actually prevent a reboot?
Those other registry entries are:
NoAutoRebootWithLoggedOnUsers = 1
AUOptions = 7
As a Software Developer, I am not comfortable deploying a change without being able to test it.