Registry Back Up and Restore?

[John John (MVP)]

Question: Have you tried it? Go ahead and try it. It can indeed be
accomplished.

Yes I have, just now again, and many times before! I have exported the
*whole* registry, *whole% hives or just select keys/subkeys and I have
then attempted to import them back into the registry. The attempted
imports had various results, either the operation failed completely
(trying to import the whole registry) or it succeeded but with errors
(when importing whole hives it imported some keys but could not import
them all) or the operation was completely successful. In all cases
there were no duplicate entries created, there were none because what
you say is not true, it is impossible to created duplicate entries in
the registry!

I have just now tested this on Windows XP and on Windows 2000 and on
both machines the results were the same. I had no qualms doing this on
production (home) machines because I have backups and I know all too
well that there can be no harm in importing identical entries in the
registry, it changes nothing in the registry. I also know that due to
the ACID nature of the registry when imports fail it will not allow the
creation of meaningless entries in the registry. In my 12+ years of
using NT operating systems I have exported/imported and merged files to
the registry countless times and I have *never* seen duplicate entries
in the registry, what you are saying is pure and simply bunk!

Anyone can easily and fairly safely test this by doing the following:

1- Create a Restore Point.

2- Create a new dummy test user account.

3- Log on to the new user account to allow the system to create the
user's profile hive. Log off to ensure that the profile hive is
properly created and stored.

4- Log back on to the new user account and use the registry editor and
export the whole HKEY_CURRENT_USER subtree, and then (while still logged
on as the same new user!) try to import or merge the exported HKCU file
back to the registry and observe the results.

As mentioned by Shenan in his post, this is akin to creating duplicate
paths or folders/files with duplicate names in the same path, it is
impossible to do! In the case of Windows Explorer, by default it will
simply assign a new, slightly different name to the file system object
(FileName (2), FileName (3)...). Other utilities may use different
methods such as adding or appending to the object name (Copy of
ObjectName, Copy2 of ObjectName...). Native Windows registry utilities
do not use these methods to rename registry entries, for starters those
newly renamed entries would be completely meaningless and useless, they
would serve absolutely no purpose. Adding or merging an entry to the
registry is an all or nothing proposition, either it works or it fails,
there is no in between and on failure the registry tools do not create
duplicate entries!

So that brings us to your claims that importing files into *your*
registry leaves it "Full of duplicates"! If this is happening then we
can only surmise or offer possibilities such as:

1- Your Windows installation is completely borked, probably brought
about by the indiscriminate use of worthless registry cleaners!

2- Your machine has a severe virus infection.

3- You are using a useless third party registry utility to import or
merge the files to the registry and the utility is using a renaming
scheme as described above to rename entries.

If, as you claim, your registry is "Full of duplicates" could you give
use the full details of some those duplicate entries? What are the
names of these duplicates?

John

 

[Twayne]

Hi Shenan Stanley,

....


Okay. I'll try anything on a computer once - particularly my virtual
machines.
Let's just do a simplistic test.

I used RegEdt32 (habit) to create the following registry values:

I don't think it matters whether it's redegit32 or regedit. IIRC I read
somewhere that regedit32 is actually a front end to start regedit.
Something I've never looked into though. Point is, you end up with the
same app running, I think.

HKEY_LOCAL_MACHINE\SOFTWARE\Twayne\Test
"TestEntryOne" (String) = "TestValue1"
"TestEntryTwo" (DWORD) = 1

I exported said values into a Registry Export file:

<copy below this line>
Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SOFTWARE\Twayne\Test]
"TestEntryOne"="TestValue1"
"TestEntryTwo"=dword:00000001
<copy above this line>

Called that file "TwayneTest.reg".

I then double-clicked on said file. The "Registry Editor" window
appeared and asked, "Are you sure you want to add the information in
C:\TwayneTest.reg" to the registry?" [Yes] [No].
....
It imported successfully now that I had permissions to it. But no
duplicate values of any sort.

Now - all I can figure is that you are doing some much larger
export/import? Although I cannot fathom how that would make any
difference. This small-scale test shows that existing values - well
- they exist and cannot be duplicated.

It's like a file system directory structure (because that is what it
is.) If I have C:\Documents and Settings\Shenan\ --> I can have that
path *once*. If the file "test.txt" exists in that directory
(C:\Documents and Settings\Shenan\test.txt) and I try to create a
file with the same name and extension there - it will tell me I
cannot.
Can you explain more fully what test would replicate what you are
claiming?
Oh - in case it matters - this was all done on a fully updated
Windows XP Professional machine - SP3, IE7, etc.

Tthat's similar to what I mean. Only the OP wanted to use Export to
export the entire registry to a file and then Import the entire registry
in order to fix a problem in it. I left his original question at the
top of this post.

I can do exactly what you did; export a key or whatever and then
reimport it, and it'll merge fine. In fact, I do it fairly often
relatively speaking. I have no idea why it's so, but importing a
complete exported registry will create the issues.
Since you're running a VM, I think you should be able to try that,
too.

Just for grins I restored my System state on the sandbox to get it
running well again, and this time imaged with XP Home SP2; the last one
was SP1, which I'd forgotten at first. It's SP2 with MSO, TurboCad and
some graphic manipulation programs, Norton, OO.o, Audacity & a bunch of
other open source stuff, AV, ZoneAlarm several spyware detectors and a
host of test utilities and a couple .NET frameworks besides a bunch of
simple 'net util apps that are mostly VB6 which don't use the registry.
The drive images are kept on an external drive, not normally connected
to the machine and it has not been connected to the 'net in the recent
past; I think two images ago. Whenever I start to use it, I re-image it
because it's sometimes left in a pretty bad state when Iv'e been playing
with things. I use Ghost14 for imaging. Two physical drives, NTFS, 80
Gig and 160 Gig.
I can only get into Safe Mode with it after Importing the entire
Registry, that was exported; it will not completely boot otherwise. Or
maybe it will, but I cut off boots at 6+ minutes. Normally it's around
3 minutes for that machine so beyond 6 minutes indicates to me something
substantially changed and historically won't repair itself with further
boots. .

I don't think the version of XP matters, nor does the Service Pack
level but I haven't actually compared results from only SP1 2 or 3. I
don't know much about, or use, VMs, but AFAIK there is very little
difference in the way an OS operates because it's in a VM.

If you have the time to fiddle, I'd be very interested in hearing what
happens when you used regedit to Export the entire registry, restart it
two or more times, or a Cold Boot, run a program or two, but not
purposely messing with the registry and then Import back in the same
Exported registry file. If I'm wrong and have some words to eat, I'll
do so, but I really feel this isn't just my particular system doing
this. This is the first time I've had anyone dispute the matter and I
have read similar experiences in other groups, though not here i don't
think.

I'm guessing and have never taken the time to check, but I think,
since the registry is constantly being rewritten as applications and the
OS are being exercised, that it'll end up with some duplicate entries
with different values to them, and with the new values actually being
the "old" ones, which can give the OS a rather strong stomach ache.
I've seen the duplicate entries and also noted that the symptoms of
importing an entire registry vary with time, and I've never ended up
with a healthy machine. But then I've only done it a handful of times,
too. I hate waiting, even for just a re-image and a boot.

I'd be interested in your own results of exporting/importing a complete
registry all in one piece.

Regards,

Twayne

 

[Twayne]

Your lead-in just doesn't sound believable; sorry. And even so, you
haven't covered the situation the OP was discussing and how/what he
would have expoted/imported vs time passing.


....

Yes I have, just now again, and many times before! I have exported
the *whole* registry, *whole% hives or just select keys/subkeys and I
have then attempted to import them back into the registry. The
attempted imports had various results, either the operation failed
completely (trying to import the whole registry) or it succeeded but
with errors (when importing whole hives it imported some keys but
could not import them all) or the operation was completely
successful. In all cases there were no duplicate entries created,
there were none because what you say is not true, it is impossible to
created duplicate entries in the registry!> I have just now tested
this on Windows XP and on Windows 2000 and on
both machines the results were the same. I had no qualms doing this
on production (home) machines because I have backups and I know all
too well that there can be no harm in importing identical entries in
the registry, it changes nothing in the registry. I also know that
due to the ACID nature of the registry when imports fail it will not
allow the creation of meaningless entries in the registry. In my 12+
years of using NT operating systems I have exported/imported and
merged files to the registry countless times and I have *never* seen
duplicate entries in the registry, what you are saying is pure and
simply bunk!
Anyone can easily and fairly safely test this by doing the following:

I wouldn't recommend this on production machines, but it should be OK.
I've only had it result in an unbootable machine one time.

1- Create a Restore Point.

2- Create a new dummy test user account.

That won't set you up to export the whole registry. You'll need Admin
privileges.

 

[John John (MVP)]

Now you're back peddling, it doesn't surprise me, that is your usual MO.
You spread FUD and when taken up on it you try to weasel your way out
of your lies.

Anyone can try this if they want, it won't hurt their Windows
installation, they can export the *whole* registry and then try to
import it again, they will get the same results as I did.

A top of doing this with Windows 2000 and Windows XP I have just now
done the same thing with Regedit on my NT 4.0 box and guess what? I got
an error message stating that I was attempting to tamper with the
registered product type, (License violation), (that is because there is
a key that cannot be touched or replaced while Windows is up and
running) and then the system continued its attempt to import the whole
file and then failed with another error message. And sure enough, on my
NT4 installation, as on my other installations, there are no duplicate
entries in the registry! Your FUD has been exposed, importing files
into the registry *never* creates duplicates!

And you haven't told us which entries were duplicated in your registry,
I wonder why? The only thing unbelievable in this whole thread is that
you still cling to your incorrect statements and still try to pass your
FUD as facts.

John

 

[Unknown]

Good grief. Did you run a registry cleaner on your brain? It's as scrambled
as your registry must be.

 

[Unknown]

You obviously have a cranium rectum inversion.

f'ups adjusted back


Nope, I am saying the same exact thing I said before. You have a very poor
handle on any of this and think you can bluff your way along; you can't.
I can't help your disability in the comprehension areas.
Using your own methodology, since you chose to not refute my suspicion
that you hadn't actually tried out what you said you did, then I'll chalk
that one up as the lie I thought it was.
There is a very simple reason why you didn't catch me in a lie too, BTW:
Unlike you, I do not lie. Period. Nothing can be gained by it, it's a
waste of time to do, and it defines the only group of people I have any
sort of actual prejudice against: Liars.



Ahh, another blatant attempt at misdirestion which you favor so strongly:
You are the first to bring NT4 into this. Shall we also discuss NT1, CP/M
and DOS operating systems like win98/95/3? Hell, you can't even figure
out which newsgroup you want to post in so you start adding your own.
That is a clear indication of what your intents actually are. But it'll
still only get read one time, if it's read at all. Even I only read them
for the entertainment value.


You didn't ask perhaps? Or maybe I have no reason to? There is nothing to
make me want to do so? You're grabbing at straws? You know and want to
see if I really do? All the above? None of the above? Since you've
actually duplicated (you claimed) the OPs propsed actions, in their
entirety, then you can easily know anyway. There was a time I'd have
answered you had you asked, but now that you've been relegated to the nut
bin, that time is well past. Especially when you can actually do it
yourself, which you have not done yet.

The only thing unbelievable in this whole

Since your allegation is untrue, than can I construe the above sentence to
mean that all else is believable? Good; not that I care. It's accurate;
believability is your problem.

You are even less believable now than you were; it seems your age
regresses with each post, actually. When I stop responding to you, you'll
know that I've relegated you to the ignorable troll bin. Be sure to hold
your breath.

Twayne