Best way to create W10 Pro reference computer and deploy the image

RBunning said:

C:\Windows\System32\bcdboot C:\Windows /l en-US<<VERY important step. Very very important- don't forget to do this! This creates the boot files. You'll get and error 'bootmgr is missing' if you forget to do this!'

How do I go about doing this part? Do I just type that line of "code" in WinPE? Or do I need to navigate there somehow?

ahelton said:

How do I go about doing this part? Do I just type that line of "code" in WinPE? Or do I need to navigate there somehow?

You just type that command. However, make sure that the drive letter at the front of the command is the letter designating the partition where the Windows image was laid down (it should be C if you did the diskpart commands in the same WinPE session and used C). Sometimes the letter is D for me and I have to type:

D:\Windows\System32\bcdboot D:\Windows /l en-US

ahelton said:

Can you further explain this:


And this since I am not doing the network approach right now:


I think I obviously must have done it correctly, but don't fully understand why I did it correctly haha.

Uh oh. Looks like I messed up when posting there- I was thinking one thing (capturing an image) but then went on to describe how to do another thing (deploy the image).

>>Rebuilding the WinPE media is only necessary when you modify WinPE's .wim, add things to the media, and if you use an .iso image instead of a USB stick (when you rebuild, it puts the new .wim and includes anything in the 'media' folder on the root of the WinPE media).

-I just want you to know about this in case you need to make a .iso or create a totally new USB stick (maybe you want to have 2 USB sticks to go around). Given you are using a USB stick for all of this, you don't need to rebuild the WinPE media each time you modify it since you can just copy your Windows image to deploy right to the USB and you can replace the WinPE .wim on the USB with a modified one (I was making .iso CD images with this for use with a virtual machine and hence described it in this way. You don't get the freedoms you do with a USB stick when working with a .iso). If you've captured your Windows image right to the USB then its on there and there's nothing special that you need to do. The next step is to deploy

RBunning said:

You just type that command. However, make sure that the drive letter at the front of the command is the letter designating the partition where the Windows image was laid down (it should be C if you did the diskpart commands in the same WinPE session and used C). Sometimes the letter is D for me and I have to type:

D:\Windows\System32\bcdboot D:\Windows /l en-US

I'm having some problems with this. Whenever I type this command in WinPE it just says "The system cannot find the path specified"

Also, how long does it typically take to:
1) Capture an image?
2) Apply an image?

I would expect it to take a bit of time since it's a multi-GB process. However, when I captured the image off the reference machine, it only took like 10 seconds. The same things occurred when I applied the image. I am a little concern that maybe something did work correctly and that's why the bcdboot process is failing. Is there a way to check whether or not the image was successfully captured?

I think I might have found the problem. On my reference computer, the system partition is assigned drive lettter C and the OS partition is E. I would assume that means I likely captured the system partition and not the partition windows is actually installed on...which explains why it didn't take very long to capture or apply. I am going to try this again and see what happens.

Yes! It worked! I've officially and successfully cloned a computer for the first time ever :) Thanks @RBunning for all the help! I guess now that my test phase is done with, I can move forward with setting my reference computer up exactly how I want it and then see about setting up some scripts to automate the process a little. I'll probably be back for questions when I get to that point. Sorry haha :)

Btw, when I redid my image capture of the reference machine it once again told me there was not enough space to capture the image. I double-checked to make sure I had my USB select as the drive to capture the image to and everything was ok there. I ended up just capturing it to my 1 TB HDD instead. That ended up working fine. I haven't checked to see what size my image is, but I'd be really shocked if it's bigger than 32GB (or really 29GB or whatever the amount of free space is on my USB). I didn't have hardly anything installed on my reference computer, so the image file should have been small relatively speaking.

A few follow up questions:

1) When you partition a drive, does that effectively treat the drive like it is really multiple drives? For example, I noticed that it was necessary to format each partition when doing this rather than just being able to partition the entire drive all in one fell swoop. I guess that would mean that you could format one partition differently than another? I understand the basics of partitioning, but certainly not some of the nitty gritty details.

2) In the Dism command to apply the image, what does the /index:1 mean?

3) Since my reference computer had the C: partition labeled as System and the E: partition as the OS, do I need to follow this same scheme for the rest of my computers that I will be applying the image to? Or is ok to set the system partition as S: and the OS partition as C: ? If found the below quote on the TechNet site and am wondering what exactly it means and why this wasn't apparently an issue for me.

On the destination computer, you will create a structure for the partitions where you apply your images. The partition structure on the destination computer must match the partition structure of the reference computer.

So I checked the size of my finalized captured image and it's only 5.77GB. Don't know why I could save it to the same USB as WinPE, since that USB is 32GB. I did come across this though and maybe it has something to do with it: clone - DISM fails with while capturing an image - Super User

ahelton said:

[...]
A few follow up questions:

1) When you partition a drive, does that effectively treat the drive like it is really multiple drives? For example, I noticed that it was necessary to format each partition when doing this rather than just being able to partition the entire drive all in one fell swoop. I guess that would mean that you could format one partition differently than another? I understand the basics of partitioning, but certainly not some of the nitty gritty details.

2) In the Dism command to apply the image, what does the /index:1 mean?

3) Since my reference computer had the C: partition labeled as System and the E: partition as the OS, do I need to follow this same scheme for the rest of my computers that I will be applying the image to? Or is ok to set the system partition as S: and the OS partition as C: ? If found the below quote on the TechNet site and am wondering what exactly it means and why this wasn't apparently an issue for me.

Great! It's so neat hearing about someone else doing this and encountering the same problem(s). That was one of the first big problems I had- mistakenly capturing the wrong partition. The next one I had was when I automated this all with a script- and I see this ties directly into your one question which I'll address in a moment. I'm glad to hear everything worked.

1>>Partitioning a drive does not necessarily mean the machine treats each partition as a whole new drive (although that's what Windows makes it look like for organizational purposes). The system and Windows sees a drive as one drive regardless of how many partitions the drive has. This is why in diskpart, when you a partition a drive each partition does not show up as an individual drive (if you type 'list disk' you'll still see the drive as one drive. If you select it and type 'list par', you'll be able to see all of the partitions on the drive). That's how the system and Windows looks at a drive.

*>>The way to look at is is partitions are parts of one disk that each have generally the same capabilities as a disk. You could even install multiple OS's to one hard disk on different partitions. Windows likely displays the different partitions as what appears to be different disks for organizational purposes.

2>>.wim images are capable of containing more than one Windows image. The /index: 1 designates which specific Windows image DISM is to apply from the .wim. You can see a list of all the indexes on a given .wim by typing DISM /get-wiminfo /wimfile:imageName.wim. Given you only just created your .wim by capturing a partition with DISM only once, there should be only one image on your .wim (you'd have to use DISM to append multiple images into one .wim).

*>>Back in the days of Windows 7, an installer's .wim would contain images of all consumer editions of Windows 7 for a given architecture. This is also the case with Windows 10 (as of today), although there are less editions of Windows 10 than there were Windows 7.

3>> I don't know what's going on in that technet quote, I would have to have a look at the whole article however I generally don't think that it has anything to do with what you're doing (I think this is a different method than what might be getting discussed in that article).

-You'll find that when you go to image computers with a custom WinPE installer (that is automated with a script), if the machines don't have the partition structure that you 'expected' it to have when you wrote the scripts the custom installer you made *might* not work... the first time you run it. (I see this happening only on the USB stick method). When you image the machines, you might notice that the script finished too early and when the machine reboots- it didn't work. If you get stuck in a situation like this, try and boot back to the WinPE media and it should work the second time around. If it doesn't- you'll need to adjust your scripts.

**> There's very high odds you might not have any problems with this, so don't stress over it.

Here's the full article on technetium where I got the quote from: https://technet.microsoft.com/en-us/library/hh824910.aspx

ahelton said:

[...}
Also, how long does it typically take to:
1) Capture an image?
2) Apply an image?
[...]

1>> It depends on where you're capturing to. I have only captured to a network server which took about 30 minutes (I was moving along at about 3 MB/s). To a USB stick, you should get about 15 MB/s, so I'd imagine the capturing would take about a third of what it was the other way.

2>>Applying my 9 GB image from the hard disk took only ~15 minutes. When applying it directly from my server, it took ~43 minutes (I don't necessarily have the fastest and greatest network setup at home for my personal use).

Exact times and speeds will depend on your setup