CloneVDI tool - Discussion & Support

[mpack]

What do you think?

I think that I would be a complete idiot if I turned down an offer like that!

By all means go ahead if this is something you want to do. You are aware of my concerns, so provided the potential problems are well signposted then I have no problem with what you are doing.

On trying the localized version, I expected you might get a kick out of seeing your words integrated into the application. I certainly get a kick out of seeing versions of my own application that I can't understand!

[MarkCranness]

Hi mpack. Have you ever thought about creating a multi-OS version of CloneVDI? Perhaps using Qt to make it portable, the same as VirtualBox does. Just wondering.

[mpack]

How does Qt work better than Wine as a multi-platform GUI? I'm not talking about programming style btw, I'm talking about advantages to users, multi-platform coverage etc.

[MarkCranness]

I haven't used Qt myself, so I don't know, sorry.

[mpack]

I haven't used Qt myself, so I don't know, sorry.

Me neither. At an early stage in CloneVDI's life I considered writing it as a multiplatform app, using gcc and the Linux native GUI API (I naively expected to find that one existed). So, I went looking for info on the native Linux GUI API, wanting to know of course how well supported it was on other platforms such as Windows and Mac. The result of my research was that there is no widely supported GUI API standard on Linux. Rather there are a plethora of the damn things, each supported by their own camps who rather than collaborate to establish a standard, prefer instead to massage their own egos and bicker among themselves about which API is best.

Except that is for Wine. Here you have a group that have taken a well established de-facto API standard and ported it to Linux and Mac platforms. The fact that they are always behind native Windows is no problem for me, because IMHO nobody who wants a wide market for their software writes it so that it only runs on the very latest Windows OS. Of course the Linux-heads snidely comment on how crappy the Win32 design is compared to their APIs. They may be right. It may also be that I don't care.

You specifically mentioned QT. I downloaded the toolkit, and that itself was a nightmare: signing up to a Nokia site followed by one of the longest, slowest downloads I've ever had to suffer. The fact that it was recently bought by Nokia doesn't fill me with enthusiasm either: IMHO that sort of thing usually happens about a year before a technology disappears altogether. And then I thought some more about it: the vast majority of PC users today are using Windows hosts, which has a native GUI API. QT support brings them no benefit whatever (it does however add an additional big download for the QT runtime). On the other hand, QT would benefit Mac and Linux users... or would it? ISTM that those users also have to download the runtime, and if they are happy to download the QT runtime then why not the Wine runtime? All in all I could see no purpose in QT, or any of the other Linux APIs.

[Sasquatch]

The only advantage Qt has is that it's cross-platform, so the program will look the same on every system. The runtime files aren't that big (well, if you compare it to your tool, they are massive), and you can even include them. But, as you said, it's not the best solution.

[mpack]

The only advantage Qt has is that it's cross-platform, so the program will look the same on every system

Isn't that true of Wine as well? Ok each platform (including the different Windows platforms) has varying themes, but I'd expect that fundamentally the UI should be the same. Though I hasten to add that since I'm not a Linux user I haven't actually checked out what CloneVDI looks like on Wine.

[Sasquatch]

Not really true for Wine. It may look the same, but XP has a nice blue title bar, Vista has it's Aero one but with Wine, the one of the system is used, Orange/brown on Ubuntu for example. Not to mention the colours of the application, they may be light grey/yellowish on XP, but 9x style grey on Wine. So, looks isn't the same all way around, it's what you select to show. But yes, the basics are the same. But that's for Qt too. Check QtGears for example, it uses the Qt toolkit and isn't that big.

[Etepetete]

Hi mpack,

Yes, I was delighted to see your GUI with my translations implemented . Then I discovered a mistake or two .
I have attached a corrected GERMAN.RC to this post.

I did a short run with the new GUI and have a question. When cloning my Slackware VDI, it recognizes the type of the Swap partition (82) but labels it as unknown. Can you give me the reason for this? By the way, CloneVDI displays the partitions in the order they appear on disk, i.e. I've a couple of Slackware VDIs and some have the swap partition at the beginning and others have it at the end of the disk.

Another thing, I'm pretty sure I formated the disks with ext4, CloneVDI is showing them as ext3. Do you know of a quick way I can check that?

And yet another question: The partition Info window shows no partition as bootable. Is this normal for Linux?

Additionally I cloned an XP VDI and it as well as the Slackware VDI ran without problems. I kept the original UUIDs and compacted the drives while cloning. It took approximately 12 - 15 minutes for each drive. It's hard to give you an exact time because the status window closes automatically when finished and I was busy correcting my spelling mistakes and couldn't keep my eye on it. Compacting worked, but as I hadn't been using the VMs much, there wasn't a large difference in VDI file size to be noted.

If you haven't noticed yet, I've gotten the first stage of the German CloneVDI FAQ up. Let's see if it brings in some more data for evaluation.

[mpack]

Yes, I was delighted to see your GUI with my translations implemented . Then I discovered a mistake or two .
I have attached a corrected GERMAN.RC to this post.

Thanks Erich, I've now integrated the new version of your resource file into my own build. Obviously I won't be releasing a new CloneVDI version for a change as minor as this, that will have to wait until I've accumulated a few more changes.

I did a short run with the new GUI and have a question. When cloning my Slackware VDI, it recognizes the type of the Swap partition (82) but labels it as unknown. Can you give me the reason for this?

Well, it's reporting the partition code from the MBR, but as it says, it doesn't recognize this code as being associated with a supported filesystem. I've now added 0x82 as a recognized code for "Linux swap". It doesn't really matter anyway, since AFAIK a Linux swap partition doesn't actually contain a filesystem anyway, meaning I can't identify unused blocks, meaning this space couldn't be compacted even if it was recognized.

Another thing, I'm pretty sure I formated the disks with ext4, CloneVDI is showing them as ext3. Do you know of a quick way I can check that?

I believe you'll find that the EXT4 format is identical to EXT3, the difference is in "soft" filesystem features that the Linux OS implements, for example journalling. This means that I can't easily distinguish between EXT3 and EXT4. Of course the difference isn't relevant to CloneVDI anyway, since the procedure for identifying used blocks is not affected by the addition of the extra features in EXT4.

And yet another question: The partition Info window shows no partition as bootable. Is this normal for Linux?

I'm no expert on what counts as normal on Linux! However, the only reason that a partition needs to be identified as bootable is because that is what the Microsoft early boot code expects. If you aren't using Microsoft boot code - for example if you are using grub - then this need not be true, though I have to say that all of the Linux bootable virtual disks that I have on my PC (including the ones that use grub) all have a standard MBR with one partition marked as bootable. On Linux this may just be a convention that isn't always followed.

If you haven't noticed yet, I've gotten the first stage of the German CloneVDI FAQ up.

Yes, I did notice. I could even read it a little bit - I was taught German at school, however that has almost entirely faded away after several decades of neglect - something I often regret. Of course if you get into technical German then I'm lost.

[Etepetete]

Another thing interests me regarding the resource file. In CLONEVDI.RC under STRINGTABLE you wrote a sub-language denominator, (SUBLANG_ENGLISH_UK). Is there a special reason for not doing the same in GERMAN.RC? (SUBLANG_GERMAN instead of SUBLANG_GERMAN_DE)

Obviously I won't be releasing a new CloneVDI version for a change as minor as this, that will have to wait until I've accumulated a few more changes.

I wouldn't have expected otherwise.

[mpack]

Another thing interests me regarding the resource file. In CLONEVDI.RC under STRINGTABLE you wrote a sub-language denominator, (SUBLANG_ENGLISH_UK). Is there a special reason for not doing the same in GERMAN.RC? (SUBLANG_GERMAN instead of SUBLANG_GERMAN_DE)

Well, I guess you could say I have a special reason, though it's a simple one! I assume you have the CloneVDI source code, so if you look at "winresrc.h" you'll see that there is no such symbol as SUBLANG_GERMAN_DE.

[Etepetete]

Hi mpack,
I remember a thread in which you mentioned that if one did a raw copy of a partition instead of the entire disk, CloneVDI would detect it and create a boot sector. So i decided to test that. I made a raw copy of the first partition of my laptop using my favorite "dumb"* imaging program. I had previously defraged but did not zero free space. I started up your fabulous tool and activated the check-box to enable compacting of the drive. Everything was functioning as I was accustomed to. Before cloning, I decided to click on the Partition Info button, expecting to see a non-bootable partition. To my surprise it showed the first partition as being bootable. Here is a screen shot of it.

RAW file partition info

PartInfo.png (9.27 KiB) Viewed 10227 times

Is CloneVDI showing me a bootable partition because it has already detected that it has to create one?

Continuing on, CloneVDI did a great job of compacting the drive. The original RAW file was 40 GB in size. The VDI was only 26 GB in size, which corresponds to the actual data present on the original partition. I created a VM with the proper settings and set the BIOS DMI data for the VM accordingly. I started the VM and ... (drum roll) ... it booted up, just as I had expected it to. So, you have another success story to add to the shine and glory of your incredible tool!

I've read in a thread in this forum, that Win7 creates a system backup in the form of a VHD, but that VHD is not bootable. Assuming that everything else to do with this kind of VHD file is equivalent to a "normal" VHD, would CloneVDI be able to detect the missing boot record and create an MBR for that kind of VHD?

Because I know you can't keep your fingers off of your code for long, here is a screen shot of the window displayed when I clicked on the Header button.

RAW file header data

Header.png (9.71 KiB) Viewed 10222 times

Note: If you take a close look at this screen shot you will see that the 'p' of 'Laufwerkstyp' (Drive type) looks as if it has been chopped off a bit.

Well that's it for now. Oh, and before I forget (again ), THANKS for the detailed responses to my questions!

*: If you haven't read this yet, here is a real nice source of information regarding 'dd'. Yes it's dumb but extremely powerful (like many a politician these days ). In the hands of an inexperienced user it is surely dangerous, but it is amazing what one can do with it, when one knows how!

[mpack]

Before cloning, I decided to click on the Partition Info button, expecting to see a non-bootable partition. To my surprise it showed the first partition as being bootable. ... Is CloneVDI showing me a bootable partition because it has already detected that it has to create one?

Yes, it has detected that the MBR (which includes the partition table) is missing, and has therefore synthesized one, with my own boot code and a new partition table as calculated by CloneVDI. What you are viewing is the partition table from the synthetic MBR. If you look at sector 0 in the sector viewer then you are again seeing a view of the new MBR. Also, if you click the ("Kopfdaten..." a.k.a "Show Header...") button then it will show you a synthesized VMDK header with two extents (the first extent being my reconstructed track 0, the second extent being the raw partition image). It shows you a VMDK header because that's how CloneVDI handles a raw file, i.e. by treating it as a flat VMDK with an implied descriptor created internally.

I started the VM and ... (drum roll) ... it booted up, just as I had expected it to.

Thanks for confirming that this feature works outside my PC!

I've read in a thread in this forum, that Win7 creates a system backup in the form of a VHD, but that VHD is not bootable. Assuming that everything else to do with this kind of VHD file is equivalent to a "normal" VHD, would CloneVDI be able to detect the missing boot record and create an MBR for that kind of VHD?

I've just had a look at that kind of VHD. In fact the backup appears to create two VHDs, a large one and a small one. I don't know what the small one is, but the large VHD does appear to be a complete disk image which already includes a valid if unusual MBR (eg. unusual in that the first sector of first partition seems to be at LBA 128 instead of the more usual 63). The only problem is that none of the partitions in the table are marked as bootable. This could easily be corrected with a hex editor, or better yet with a partitioning tool running inside a VM (obviously the VHD would have to be mounted as a second drive first). I'm not sure that CloneVDI is the best tool for specialized manipulations like that.

Note: If you take a close look at this screen shot you will see that the 'p' of 'Laufwerkstyp' (Drive type) looks as if it has been chopped off a bit.

Thanks for pointing that out, now fixed in my copy. I thought I had caught all of those!

[Etepetete]

Hi Don,

I've added another batch of translated info to the German FAQ. This "head banging" has proven good for me, as I have had to scrutinize your text in detail, which allows me to learn a great deal. I suppose you know how far along I've gotten, based on the questions I've been asking you (hint: RAW + MBR).

Regarding the translation of the word 'extent': I don't want to get into discussions about technical jargon and whether or not it should be translated or not, but after having researched extensively I have not found any translation that would apply. Considering that your tool is more for the not so technically versed, I scoured my brain for an explanation of 'extent' and it came to mind, that I could translate the word into 'Datenbereich' (Data area). If you agree that this would be a suitable translation, then please update the resource file accordingly. (I hope it isn't necessary for me to upload a modified resource file for this one change.)

Should you make changes to your release notes, it would be much appreciated, if you could just pass me a hint as to what changes you made, so that I don't have to go through the whole file, checking line for line against my translation. (You know how much I hate this kind of work.)

Thanks in advance,
Erich