321 replies to this topic

[sambul61]

Did you mean, the brown icon shows, how many fragments a part of a VHD file housing a selected volume is comprised of, when pointed at by a Magnifier Glass? I think, contiguity of a VHD file should be reflected on Hard Disk Panel, while info about VHD file content - on VHD Disk Panel. The panels can be selectable as Tabs.

Btw, a filedisk must be contiguous to be mapped by Grub4DOS, but once mapped, I'm not sure, if its volumes must be contiguous too - needs to be clarified. On a hard drive its always seems to be the case - the volumes are always contiguous.

As to depicting fragmentation in iPhone style bar, I'm not sure that mixing styles signals the best taste , especially in a tool targeting primarily computer users, but not mobile devices. May be its better to stick to the same style, as in the lower bar (or in my plagiarized pictures)?

[sfinktah]

Just a mockup.. and the fragmented icon merely shows that there is fragmentation. More information could be obtain on click/hover. For G4D, the VHD file would have to be contiguous. Not sure if the files stored inside have to be, they'd probably follow the same restrictions as other image based boot sources.

The green and gold blocks would replace the existing partition panels on the left. The reason there are two identicons shown, is that it will be a requirement for Differencing VHDs, which will need to identify their parent.

[sfinktah]

on VHD Disk Panel. The panels can be selectable as Tabs.

Interesting idea... funnily enough, I've already written a "Search" function that check the root paths (eg, c:vhd, c:vhds, but not any subdirs beyond that) for vhds.

but you just can't display them as nicely as you would like... if they're fragmented to heck, it's going to just be impossible. Unless you wanted to see something like

[sambul61]

I think, the existing volume depictions are fine. Too much color destructs attention from content, while too complex visualization would prompt an untrained user to take some headache pills. The right proportion of text and visual content works wonders.

[sfinktah]

Too late. And I wouldn't worry about the number of colors, it's unlikely that you'd ever see any color other than green (NTFS). Unless we decide to depict unused space as a partition - but that would be grey anyway.

Bloody photoshop. Somewhere between Photoshop on my Mac, and my development PC, it totally screws my colors.

[sambul61]

Using VHDMount certainly is a totally different way to mount a VHD, and totally different backend drivers. But I like it - it has an inbuilt "commit" flag when you close the VHD, just like mounting a WIM.

I don't really think it's the kind of driver that would ever work as "bootable" anyway.

Yet another way to boot from VHD an OS version unsupported for native boot is VBoot. Its author also distributes VBoot Port Driver, which is similar in fuction to WinVBlock and FiraDisk, but supports hot-swapping filedisks mapped by VBoot. When looked at closely, it might give some extra ideas on booting non-contiguous images as well, in addition to hot-swapping techniques.

Some holiday notes.

[sfinktah]

not so bad, eh?

[sfinktah]

Yet another way to boot from VHD an OS version unsupported for native boot is VBoot. Its author also distributes VBoot Port Driver, which is similar in fuction to WinVBlock and FiraDisk, but supports hot-swapping filedisks mapped by VBoot. When looked at closely, it might give some extra ideas on booting non-contiguous images as well, in addition to hot-swapping techniques.

have you actually tried to download that lately? ... or was it some-thing else that refused to download without registration...

Anyways, I noticed that 2008 R2 has two (visible) driver components... apart from the oft-seen


There is the oft-overlooked (but what I think could be more important boot-wise for your cheap-Windows-7 versions):



It would be interesting to compare all the options (there are so many) set for both, and compare.

[sambul61]

I feel your green being too intense for my taste... And VHD "inside space total fragmentation bar" is missing that indicates each volume zone fragmentation (if they become fragmented during a dynamic VHD grows - that's a separate question to investigate), and how full the VHD is with data & space allocated for volumes. Disk Management allows to create only Basic volumes on a VHD, but what about Diskpart? Also, when a Basic Volume is created inside a dynamic VHD, does that mean, its full size is preallocated immediately as contiguous space?

Being a member of the Vboot forum, had no problems with downloads. That forum may be the right place (and it was) to ask about Vboot particulars and offer suggestions.

I think this SCSI adapter starts as service at boot time, thus allowing to actively search for VHDs mapped by Bootlrd as SCSI drives. WinVBlock seems to work in similar fashion.

P.S. I created a dynamic VHD, then added 2 simple volumes to it, then copied a file to 2nd volume, them another large file to 1st volume. The volumes were not fully allocated when created, but growing with the VHD. Space inside each remains non-fragmented as per PerfectDisk, and the VHD file contiguous - didn't test it in a limited space yet. I wonder, how to find out, if the space occupied by volumes remains contiguous inside the VHD. Since the 2nd volume was enlarged 1st by adding a file to it, and then 1st volume was enlarged by adding another file to it, but it didn't result in 2nd volume be moved to the right before 1st volume was allowed to grow (at least no time was spend on it), the volume allocation inside a dynamic VHD remains a mystery thus far.

[sfinktah]

P.S. I created a dynamic VHD, then added 2 simple volumes to it, then copied a file to 2nd volume, them another large file to 1st volume.

How did you copy it though?

As for the green, it certainly wasn't the color I chose, but I have a color picker in the debug section, you can have a go at a nicer shade.

And VHD "inside space total fragmentation bar" is missing that indicates each volume zone fragmentation

That wasn't really showing anything useful. It was basically a mirror of the partition cluster usage, plus any "waiting for compaction" extra space. It was adding little value, and making the UI messy. I'll end up placing that in a dedicated view.

(if they become fragmented during a dynamic VHD grows - that's a separate question to investigate), and how full the VHD is with data & space allocated for volumes. Disk Management allows to create only Basic volumes on a VHD, but what about Diskpart?

Can't say I've attempted to test it, just going by what I read on technet:

Understanding Virtual Hard Disks with Native Boot
Published: October 22, 2009
Updated: October 22, 2009
Applies To: Windows 7, Windows Server 2008 R2

Native VHD support has the following limitations:

• Native VHD boot is supported only by Windows 7, and it is restricted to the following editions:
  
  • Windows® 7 Enterprise
  • Windows® 7 Ultimate
  • Windows Server 2008 R2
• Native VHD boot does not support hibernation of the system, although sleep mode is supported.
• An attached VHD cannot be configured as a dynamic disk. A dynamic disk provides features that basic disks do not, such as the ability to create volumes that span multiple disks (spanned and striped volumes), and the ability to create fault-tolerant volumes (mirrored and RAID-5 volumes). All volumes on dynamic disks are known as dynamic volumes.
• The parent volume of the VHD cannot be configured as a dynamic disk.

[sambul61]

Interesting, just decided to look what's going on. Simply attached the VHD, assigned drive letters to its volumes, and copied files to the volumes like ordinary drives. As to the cited MS doc, many such docs were proven incomplete or in part wrong by some developers of this and other similar Boards, as they didn't go that far as some curious user minds. Nonetheless, studying NTFS Virtual Volumes behavior created inside VHDs is waiting for some explorers.

What happen to your "Pretending Signature" line?

New Year is coming to Australia pretty soon. According to this, now its 3pm Dec.31, 2011 in the capital, and OMG +31 ⁰C, sounds like a nice winter with plenty of snow and even some ice - watch your steps. We are way behind... Hope, we are going to meet 1st VHD Director pre-alpha soon...

Happy New Year!

[sfinktah]

Haha, well, stop distracting me.



And if that isn't enough, feel free to modify the PNG... it's only a few lines of code away from being completely skinnable anyway.

Attached Files

•  GreenPartitionFill.png   3.77KB   1 downloads

[sambul61]

Nice tute... It occurs to me, that this procedure could be make fairly automatic, if someone (me, I guess) were to write a small application that ran under WinPE.

I gather it's not "legal" to distribute a modified copy of WinPE? A pity, as one could just release a pre-build WinPE with everything on board. (I run my own from PXE, works quite well).

That said, it also shouldn't be hard to write an application to make a WinPE... it all depends on whether the application can get hold of those few tools it needs, without having to grab the whole AIK ISO.

Wow... a skinnable GUI. It might be the right time though to frozen GUI development, until Alpha shows up. So, the next step might be... adding an in-house Install OS Plugin that would serve as a sample plugin, while automating Win7 install from a downloaded by a user Win7 Setup ISO to a VHD as planned?

VHD Director would use that Plugin to do the following:

- use Http Driver to selectively download ImageX from WAIK package,
- create a VHD, then apply a suitable WIM from the ISO to VHD
- inject WinVBlock (for unsupported for native boot OS versions) as per Install Unsupported OS to VHD Tutorial (Step 4, alternative method)
- add USB Boot support as well to the offline OS on VHD
- add BCD boot environment to the VHD and to a hooked to that PC empty properly pre-formatted by a user USB Thumb (if any)
- add Grub4DOS files to the Thumb and G4D Menu entry to its BCD
- activate Test Mode (for 64-bit unsupported for native boot OS with WinVBlock installed)
- copy the VHD file to the USB Thumb
- prompt user to adjust BIOS boot order and reboot OS from VHD on the Thumb

It seems, our prospective users are hungry by now for a "real" Alpha deal that would do what no other VHD tool does. And that would prompt development of the app Plugins, and also much broader usage of VHDs initially on this Board for experimenting with running OS from VHD, replacing WinPE with componentized & customizable WES7, and such... In turn giving plenty of feedback and feature requests for VHD Director continued development. Ordinary users are really hungry now for Easy VHD Button...you so responsibly promised. And it seems fairly easy routing to automate first without discussed earlier VHD manipulation complexities.

And may be you can look at my PM at some remote point in time...

[sambul61]

Can't say I've attempted to test it (dynamic volumes), just going by what I read on technet:

Understanding Virtual Hard Disks with Native Boot:

An attached VHD cannot be configured as a dynamic disk. A dynamic disk provides features that basic disks do not, such as the ability to create volumes that span multiple disks (spanned and striped volumes), and the ability to create fault-tolerant volumes (mirrored and RAID-5 volumes).

Interesting, because I read a different thing on technet, posted on the same date:

"A dynamically expanding VHD is a physical disk that you have initialized for dynamic storage. It contains dynamic volumes such as simple, spanned, striped, or mirrored volumes or RAID-5 volumes."

Also, blocking Hibernation support in OS booted from VHD seems to be marketing related, as such feature is also blocked by default in Win Server, but can be unblocked by a curious mind. Will try with VHD...

Btw, while reading it, come up with a new feature:
- start VHD Director at boot time, and auto mount the VHDs, selected by a user or attached before last PC shutdown.

[sfinktah]

Interesting, because I read a different thing on technet, posted on the same date:

"A dynamically expanding VHD is a physical disk that you have initialized for dynamic storage. It contains dynamic volumes such as simple, spanned, striped, or mirrored volumes or RAID-5 volumes."

That's from the same article, and you're quoting it out of context. It's much funnier in context.

"You cannot configure two attached VHDs to be a dynamically expanding VHD. A dynamically expanding VHD is a physical disk that you have initialized for dynamic storage. It contains dynamic volumes such as simple, spanned, striped, or mirrored volumes or RAID-5 volumes."

It looks like it has been written by someone who had an absolutely empty brain, and has accidentally confused "dynamic volume" with "dynamic virtual disk" while consulting some kind of glossary.

[sfinktah]

But in defense of your argument, the same document also states:

"Booting to a VHD that is located on a remote share or a USB flash drive. Windows does not support booting to a remote share or a USB flash drive, whether installed on a physical volume or from a VHD. You can boot the Windows Preinstallation Environment (Windows PE) from a USB flash drive, which is supported for Windows deployment. Windows PE typically boots from either a Boot.wim or an installed image, but booting Windows PE from a VHD is not supported."

And we know that's nonsense.

Feel free to create a dynamic volume in a VM and then mount it standalone, and let me know how you go.

[sambul61]

That's from the same article, and you're quoting it out of context.

It looks like it has been written by someone who had an absolutely empty brain

Except your above link points to Understanding Virtual Hard Disks with Native Boot article, and mine to FAQ: Virtual Hard Disks in Windows 7 and Windows Server 2008 R2 . May be they were written by different Technet Doc writers, who possibly consulted different MS teams or members, while they were discussing overtime how this feature would evolve. Also, booting from a removable USB drive was not officially supported in any Windows version at the time of publication.

[sfinktah]

Sorry - upon closer inspection, they are different articles. But not by much. A particularly cogent case being the quote currently being bandied about.

"You cannot configure two attached VHDs to be a dynamically expanding VHD. A dynamically expanding VHD is a physical disk that you have initialized for dynamic storage. It contains dynamic volumes such as simple, spanned, striped, or mirrored volumes or RAID-5 volumes."

The matching sentence from the earlier link:

"An attached VHD cannot be configured as a dynamic disk. A dynamic disk provides features that basic disks do not, such as the ability to create volumes that span multiple disks (spanned and striped volumes), and the ability to create fault-tolerant volumes (mirrored and RAID-5 volumes). All volumes on dynamic disks are known as dynamic volumes."

Spooky, isn't it. But the latter version (Understanding Virtual Hard Disks with Native Boot) is clearly the correct one. (Publishing date is actally 1 week after the other article Frequently Asked Questions: Virtual Hard Disks in Windows 7 and Windows Server 2008 R2)

[sfinktah]

I liked it so much I used it as text for the first Wiki entry.

http://sfinktah.trac...iki/native_boot

[sambul61]

But the latter version (Understanding Virtual Hard Disks with Native Boot) is clearly the correct one. (Publishing date is actually 1 week after the other article Frequently Asked Questions: Virtual Hard Disks in Windows 7 and Windows Server 2008 R2)

Actually, the article linked by me was last updated on Oct.15, 2010, while linked by you - on Oct.22, 2009. Does it say anything to you? I don't see any relevant close matching sentences in these articles as well...

Btw, just a day or two ago, when I looked through it, the last update dates were the same. Is seems MS guys are reading this thread, and fixing their "typos" on the fly - not always adequately though... You might want to reconsider your Wiki article choice...given recent discoveries.

Speaking of test proofing the MS induced wonder, in my initial non-complete tests the NTFS Volumes created on an attached dynamic MBR-based VHD disk behaved more like dynamic non-contiguous volumes rather than basic contiguous ones. They didn't span though over several physical disks, but may be because I didn't make the dynamic VHD big enough to span over several physical disks. Even if they can't span, there behavior needs further investigation, if we want to deal with them in a controlled fashion.

[sfinktah]

Actually, the article linked by me was last updated on Oct.15, 2010, while linked by you - on Oct.22, 2009. Does it say anything to you? I don't see any relevant close matching sentences in these articles as well...

Btw, just a day or two ago, when I looked through it, the last update dates were the same. Is seems MS guys are reading this thread, and fixing their "typos" on the fly - not always adequately though... You might want to reconsider your Wiki article choice...given recent discoveries.

Oh, didn't notice the year. Silly me. Why reconsider? My article is still the correct one. I guess they're not excessively similar, obviously any VHD article is going to cover certain similar concepts, might be interesting to make a list, there's stuff missing from both... might be able to actually produce a decent document

eg, mine doesn't mention: "You must have volume management privileges (which is granted by default only to administrators) to attach a VHD because attaching a VHD is equivalent to bringing a hard disk drive or volume online."
and yours doesn't mention "Native VHD disk management support can attach approximately 512 VHD files concurrently."

Mine says "VHD files cannot be nested."
Yours says "You can only attach two nested VHDs. When you create a VHD within another VHD, it is referred to as a nested VHD. The limit for nested VHDs is two. That is, you can attach a VHD within another attached VHD, but you cannot attach a third."

Mine says: "The local disk partition that contains the VHD file must have enough free disk space for expanding a dynamic VHD to its maximum size and for the page file created when booting the VHD. The page file is created outside of the VHD file, unlike in the case of a virtual machine where the page file is contained inside the VHD."

Yours says: "Ensure that there is sufficient space on the host volume for paging files (Pagefile.sys). During native boot, a dynamically expanding VHD is automatically expanded to the maximum size on the host volume, and the paging file is created on the host volume outside the virtual volume. The paging files must be located on a physical volume outside the VHD for system performance. If the host volume does not have enough free space for a paging file, Windows attempts to find free space on another volume. The paging file size depends on how much physical RAM is available on the system (you should estimate approximately 5 GB of available space in addition to the maximum size of the VHD file)."

Mine: "The local disk must have at least two partitions: a system partition that contains the Windows 7 boot-environment files and Boot Configuration Data (BCD) store, and a partition to store the VHD file. For more information about disk partitions, see Understanding Disk Partitions. For more information about adding a Windows 7 boot environment for native VHD boot, see Add a Native-Boot Virtual Hard Disk to the Boot Menu."

Yours: "Run Sysprep to generalize the image before using a VHD for native boot on a different computer. Sysprep generalize prepares a Windows image that is installed on a physical partition or on a native boot VHD that is to be used on another computer. After you run Sysprep, you can copy the VHD to multiple physical computers or virtual machines for native boot. During the first boot from the VHD, Windows completes the specialize configuration pass to detect the hardware devices and initializes Windows to run on the new computer."



I guess that's just proof of what you say "way back when" - there are a lot of article (even microsoft ones) - that simply don't agree.

Speaking of test proofing the MS induced wonder, in my initial non-complete tests the NTFS Volumes created on an attached dynamic MBR-based VHD disk behaved more like dynamic non-contiguous volumes rather than basic contiguous ones. They didn't span though over several physical disks, but may be because I didn't make the dynamic VHD big enough to span over several physical disks. Even if they can't span, there behavior needs further investigation, if we want to deal with them in a controlled fashion.

Now you just totally lost me, and I even followed your hyperlink and re-read the description of the original tests. And it's not helping that the last sentence (which I have placed in italics) only appears in my [quote] but not in your original article.

Very very very strange.

[sfinktah]

Speaking of test proofing the MS induced wonder, in my initial non-complete tests the NTFS Volumes created on an attached dynamic MBR-based VHD disk behaved more like dynamic non-contiguous volumes rather than basic contiguous ones. They didn't span though over several physical disks, but may be because I didn't make the dynamic VHD big enough to span over several physical disks. Even if they can't span, there behavior needs further investigation, if we want to deal with them in a controlled fashion.

Speaking of test proofing the MS induced wonder, in my initial non-complete tests the NTFS Volumes created on an attached dynamic MBR-based VHD disk behaved more like dynamic non-contiguous volumes rather than basic contiguous ones.

screenshot says:


Where did that mystery sentence come from?

[sfinktah]

New Year is coming to Australia pretty soon. According to this, now its 3pm Dec.31, 2011 in the capital, and OMG +31 ⁰C, sounds like a nice winter with plenty of snow and even some ice - watch your steps.

Oh dude, you have no idea. It was officially 103ºF when we went for a drive at 6pm (Jan 2nd) or 106ºF if you prefer to trust BMW.

That's "toasted bare feet on asphalt" weather. 3 days later, and it's practically the south pole.

I was in DC until 17th November last year, which was pretty much perfect timing temperature wise. Just when it was getting cold enough to gas up the fire, I was back in Oz wearing the shorts and t-shirt I was wearing 3 months previously in the states.

[sambul61]

The page file is created outside of the VHD file, unlike in the case of a virtual machine where the page file is contained inside the VHD

I guess, Windows placing Pagefile outside of the VHD might be connected to a fixed contiguous VHD becoming non-contiguous after booting inside a VM... It might also imply, other NTFS service files critical for that volume are also placed outside of the VHD after it boots inside a VM, thus making it reported as non-contiguous by Grub4DOS at next reboot. It doesn't seem to happen though, when there's enough free space left on a fixed contiguous VHD to accommodate all these files. Also, It seems to only happen, when the VHD is booted inside a VM as a filedisk stored on another VHD, rather than a virtual hard drive directly connected to the VM.

[sfinktah]

Having only thought about it for the time required to read your post, I can see no reason to place a pagefile outside of a fixed vhd other than "it would be faster."

There aren't many more compelling reasons to be said for dynamic vhds.

But that's when looking at it from a "single user, single copy" perspective.

If you look at it from a "distributed/iscsi" enterprise angle, you might find more compelling reasons for people to keep their dirty little swap files on their own PCs.

Ultimately a swap file is simply an extension of physical memory and is only stored to disk because that happens to be the second fastest thing laying around. So there is certainly no reason it should be stored in the same place that stores persistent (between reboot) data if there is any alternative.

Unix based systems (including Macintosh I believe) store their swap files on separate partitions that are reserved purely for swap.