Thank you very much :)
I am so glad that you have asked! I absolutely never mind explaining what I'm doing! If the below explanation doesn't make complete sense, please do quiz me on it - I positively encourage you to ask questions about it :)
Right, we're looking at a
repeated reboot scenario. Something somewhere's gone wrong: fact. First, let us think about an overview of the Windows Update mechanism for an update which needs a reboot.
Windows Update detects and downloads update.
Initial extraction and installation occurs.
Reboot required flag set.
Windows Update asks for a reboot.
User reboots.
Reboot notices the flag, performs latter part of pended installation, and unsets the reboot flag.
Done :)
But what if something went wrong and the reboot flag was never unset? Well, the reboot would repeatedly occur. The reasons why it doesn't just get unset on the next reboot are complicated and vary greatly case to case. For example, if one part of the pended installation got stuck, it could also flag a reboot even without the reboot flag being set. And as it's "stuck", it's repeated.
Well, as it happens, there are many, many files and registry keys/values which can cause a reboot to be demanded. And the point is, things aren't quite in order: this is not a case of everything being set properly, this is a case of one flag being stuck.
But the problem is, I don't know which one it is. So I have to keep looking in different registry locations until I find it.
The first set of commands I set were to clear out any stuck fragment of a larger installation. These did no good, so this isn't a case of a stuck fragment: instead, we have a registry flag which is stuck at set.
So I took exports of two common flags to see if they were set, and I came across something rather curious. Yours is not the simple case (alas, computers can never be simple, it seems!): you don't have a single stuck reboot flag. You have a positive explosion of pended tasks and reboots flags.
I will need some more time to consider this. Care must be taken here, and a hasty move should not be made. I must consider carefully what is going on, and I may need to take some additional registry exports to see how deeply rooted the problem really is.
What are your questions?
Richard
