Rationale for Deleting Restore Points after Windows Update

Hi, I believe you are referring to a feature update, when all system files are replaced. The changes are so significant that a system restore point would be
a. Utterly wrong to try to apply
b. Cause instability
c. Result in you having a mix of system files from the old build and the new
d. Make registry changes that might not be relevant to the new build
etc.

The common advice is that before any major update

- that's only applicable to a given build of the O/S. Further, restoring is notoriously unreliable- if it works, fine, if it doesn't, you'll just be back as you were, no harm done.

System restore points can also undo any Windows updates applied during the relevant period. What meaning would there be in undoing an update applied during the old build in the context of the new?

You would be hoping that using a restore point would cause all new system files to be replaced by old ones. That can't happen as only changes in a given build are protected.

The more you think about it the more inconsistencies occur that could be introduced were a restore point used as you would like. Consider, for example, the implications of using offline DISM after having used such a restore point.. the inconsistency introduced by using such a restore point with respect to the WinSxS database.

In short, there is no way whatsoever a system restore point is relevant to or usable in the context of wishing to revert to the precious build.

There is a completely different mechanism provided - 'Go back to the previous build' - which is available for 10 days by default after upgrading. Now you'll rightly say that's fine, provided the upgrade works normallly. What if it fails badly?

How you should routinely protect your PC
The way you protect yourself and your PC from this and many more difficult problems, including disk failure, is to routinely use disk imaging. (E.g. Macrium Reflect - free/paid).

mrhagerty said:

A much older thread discussed the reasons manual RPs get deleted, one of which was that the older restore points wouldn't apply to the new Windows version just updated.

The logic seems to overlook THE most important reason one would want a restore point available after an update - if that update fails or updates incompletely.

The common advice is that before any major update one should make a restore point in case something messes up. Yet Windows logic takes that option away at the most needful of times - when one is counting on getting back to a successful Windows version when something goes awry.

What I'm missing is why you can't restore a state of Windows before a version update, since it replaces the system just erroneously installed?

Is this a case that system restore doesn't backup all system files from the current version and depends on certain files of the old version to be there? It just seems incongruous that the most important need for a restore point is defeated by "the way Windows file dependencies work."

Can someone explain?

RP is not a backup, never was, never meant to be a backup. Other, better options are available.

Unfortunately they are not equivalent- restoring a partition means restoring everything- including data. One very good reason for keeping as much personal data off C: as possible.

Perhaps the nearest thing to SR is Rollback RX (free version is adequate, apparently) - but I'm uncertain about having it active when upgrading- may need to remove it.

Steve C said:

I'm so impressed with Macrium Reflect that I've disabled System Restore!

I recently decided to have one or two SRPs, rotated weekly or bi-weekly because for some minor implosions, explosions, an SRP rather than a total OS restore may be all that's needed.

dalchina - would the known good configuration found within the windows 7's F8 menu be updated by a current SRP, or, where does that known good configuration come from?

RolandJS said:

dalchina - would the known good configuration found within the windows 7's F8 menu be updated by a current SRP, or, where does that known good configuration come from?

Sorry, I don't know- as you would assume, it depends where the data representing that is held, and whether that is included in what SR protects (which is not clearly defined as far as I know). I would suggest only research and experimenting would tell you, unless you are lucky enough to find a relevant informed comment on this.