46 replies to this topic

[r4ymonf]

Hi! I... uh... installed Windows XP onto a VHD with VirtualBox following this tutorial. I put it onto a flash drive but I'm getting Error 60! I already tried WinContig, and then it failed with "The volume does not have a group of contigous, free clusters that is large enough to contain the entire file.", I've tried changing the cluster size from default to 64K and I've tried 4 times already. I'd really appreciate some help. Thanks.

[r4ymonf]

P.S. I've tried NTFS and FAT32, both are the same. [Same error in wincontig]

[Wonko the Sane]

Basically what you are experiencing is that you have (either before you started fiddling with it or because of your fiddling with it) such a heavily fragmented filesystem on the USB stick and/or not enough "spare" contiguous space that Wincontig has not enough "workspace".

Cluster size or filesystems used (in the .vhd) is totally irrelevant.

You can either "start form scratch" after having re-formatted the USB stick or use a "complete disk" defragment program on it (before copying to it the .vhd).

HOW exactly (under which OS and through which specific tool) you are copying the .vhd to the stick may be part of the issue.

Additionally how BIG is the resulting .vhd?

Which size is the USB stick?

 

Wonko

[r4ymonf]

What program can I use for a "complete disk" defragment?

1] TeraCopy Pro and the included windows copier

2] About 6GB

3] 8GB - 7.4GB recognised

[Wonko the Sane]

Good so you have no issues with Ram (you have plenty of it).

You can use either a third party such as UltraDefrag:

http://ultradefrag.s...t/en/index.html

or the Windows built-in defragmenter.

The issue is that a USB stick will normally be seen as Removable and normal Windows may have issues with it.

You can then use a filter driver:

http://www.msfn.org/...os/page__st__54

to have it be seen as "Fixed". (I normally flip the bit directly on the hardware but it is an operation that is not "easy").

 

Wonko

[r4ymonf]

Okay, I'll try that. Thanks. I wouldn't even know how to start opening a cruzer edge flash drive.
Oh noes, the link is dead for the filter I believe.

Oh, and I formatted everything and then I defragged but there were no files.

Edited by r4ymonf, 19 January 2013 - 03:15 PM.

[Wonko the Sane]

Well, there is no need to "physically open" the thingy.
WHICH link is dead?

Any problem with the meaning of "either/or"?


Wonko

[Sha0]

Oh, and I formatted everything and then I defragged but there were no files.

The simplest way to have the best chance of success with a contiguous file is to backup all files from the target, reformat the target, copy only the desired file back to the target, run a tool to make it contiguous, then restore the rest of the files from the backup.

[Wonko the Sane]

The simplest way to have the best chance of success with a contiguous file is to backup all files from the target, reformat the target, copy only the desired file back to the target, run a tool to make it contiguous, then restore the rest of the files from the backup.

Which BTW, expecially if the file that is copied is a "largish" one is also the perfect recipe to hit a CHS limit for loading the system files such as NTLDR or IO.SYS and more generally all kind of issues with limited LBA access from bootsector code.

 

JFYI:

https://www.virtualbox.org/ticket/959

 

Seriously, the "old school" way is to copy to a newly formatted disk ALL loaders/system files that may be needed/used in a boot environment, and only later copy the largish .VHD file that needs to be kept defragmented.

 

Wonko

[r4ymonf]

I don't understand, should I do the "old-school" way (which I have no idea on how that'd work because I've already tried that ) or Sha0's way?

[Wonko the Sane]

I don't understand, should I do the "old-school" way (which I have no idea on how that'd work because I've already tried that ) or Sha0's way?

It depends on how big is the .vhd image and how big is the device and how (with which geometry) the BIOS recognize it.

Normally there won't be any difference, as the USB stick (which is roughly 8 Gb in size) will be detected as having a H/S geometry of 255/63 and the LBA limit for 255/63 is actually around 8 Gb.

Your .vhd is surely smaller than 8 Gb (around 6 Gb) so you should have no issues using either approach.

Simply copy only the .vhd (i.e. do NOT multi-select a number of files) and check with wincontig if it is "delivered unfragmented".

Report what happens.

 

Wonko

[Sha0]

Which BTW, expecially if the file that is copied is a "largish" one is also the perfect recipe to hit a CHS limit for loading the system files such as NTLDR or IO.SYS and more generally all kind of issues with limited LBA access from bootsector code.

 

JFYI:

https://www.virtualbox.org/ticket/959

 

Seriously, the "old school" way is to copy to a newly formatted disk ALL loaders/system files that may be needed/used in a boot environment, and only later copy the largish .VHD file that needs to be kept defragmented.

While I agree that your post is true and useful to mention for this scenario, it stands on its own and has no bearing on the truth of what I said.  I always use the precise method you've described as "old school," but that is not relevant to the problem of making a contiguous file.

 

I have a problem with your post, and that is that it could have gone like this, instead:

And in this case, since you will be needing NTLDR, IO.SYS or other system files to be at a "low place" on the disk to prevent C/H/S issues, you will want to copy just those files first, then the desired "big" file, then make the big file contiguous, then copy the rest of the files.

Instead, my post is portrayed as "a perfect recipe [for issues]", when in fact, it's perfectly valid as-is.  You can argue that it is irresponsible to "mislead" someone with information that, while true, might not be the whole story, but I'd argue that lessons are sometimes best learned by experience ("NTLDR is missing") than by theory.

[Wonko the Sane]

@Sha0

Besides the fact that I was (ad still am) joking, you have to admit  that the thread is not about "making a file contiguous" but rather about "making contiguous a large .vhd file that I want to boot following this tutorial http://www.linuxbyex...n-usb-disk.html BUT I have an error about the file not being contiguous".

I am looking at the wider scenario .

 

Wonko

[r4ymonf]

I:

1) Formatted to NTFS and then checked for any fragmented files.

2) Copied the VHD only.

3) IT'S FRAGMENTED AND SAME ERROR [in grub4dos and wincontig].

D;<

Grub is mean.

[Sha0]

I:

1) Formatted to NTFS and then checked for any fragmented files.

2) Copied the VHD only.

3) IT'S FRAGMENTED AND SAME ERROR [in grub4dos and wincontig].

D;<

Grub is mean.

Then you don't have enough space.

[Wonko the Sane]

Hmmm.

A 6 Gb .vhd is very big for an XP install, but still, it shouldn't get fragmented during copy operation.

 

Is the "original" (on the hard disk) fragmented or contiguous?

 

HOW exactly are you copying the .vhd?

 

Are you running an antivirus or some form of "online/background" file scanning?

 

Idea

How many fragments do you get?

If they are just a few it is possible that the issue is the $MFT (and/or some other NTFS system files) that break the image.

 

In theory it should be possible to move the $MFT and all other system files to the beginning of the partition, though right now I can remember no "automated" tool capable of doing it. 

 

I need the EXACT size of that stick and partition (in bytes).

Get Hdhacker:

http://dimio.altervista.org/eng/

make a backup of BOTH the first sector of the Physicaldrive and of the first sector of the Logicaldrive.

Compress them together to a .zip file, upload the .zip to *any* free file hosting service and post a link to it.

 

I can try making manually a suitable "blank" filesystem for your stick this way.

 

 

Wonko

 

 

[r4ymonf]

1) The original seems to be fragmented even though I've defragged it from VBox a LOT of times.

2) I'm using TeraCopy Pro.

3) Avast, I had that off while I was copying because it was conflicting with Trickster Online [I was playing that while I was copying].

4) 2 fragments.

5) Um, I don't know how to add an attachment, so: mediafire link here.

[steve6375]

It would be interesting to know the size of each fragment (and position). If you can copy the vhd over and then install grub4dos, boot to grub4dos to the command console, and then type

blocklist /my6gb.vhd

and report what you get. Then repeat the format, copy VHD, add grub4dos, do blocklist command again and see if you get the same size for each of the 2 fragments.

 

 

If the size of each fragment varies, then you have some sort of 'random' process which is causing the file to be split.

If the fragments are the exact same size each time, then you have a repeatable, deterministic problem. As it occurs on both FAT32 and NTFS then it may be some other process.

 

 

TeraCopy uses dynamically adjusted buffers to reduce the seek times. Asynchronous copying speeds up the file transfer between two physical hard drives.

 

TeraCopy Pro may actually split the copy into two parts in some way to speed it up (e.g. make a sparse file of the same size as the vhd, then copy from beginning to middle and from middle to end - at the same time, asynchronously) - can you just use normal OS copy instead?

 

Also try the vhd copy in safe mode (and not doing anything else whilst you are waiting)?

 

P.S. In post #2 you said you tried copying the VHD to both FAT32 and NTFS - but if the VHD is 6GB, how did you copy a 6GB file to a FAT32 volume (the max filesize is 4GB)?

[r4ymonf]

P.S. In post #2 you said you tried copying the VHD to both FAT32 and NTFS - but if the VHD is 6GB, how did you copy a 6GB file to a FAT32 volume (the max filesize is 4GB)?

I actually made the VHD smaller, as I had 2 versions of the VHD: 3.9GB and 6GB.

 

TeraCopy Pro may actually split the copy into two parts in some way to speed it up (e.g. make a sparse file of the same size as the vhd, then copy from beginning to middle and from middle to end - at the same time, asynchronously) - can you just use normal OS copy instead?

Oh, I never noticed that. I'll try that, thanks!

[r4ymonf]

Still didn't work, perhaps it's because it's fragged on my disk, and WinContig gives me the same error on my C:\.

[steve6375]

No, source should not matter.

Can you try running blocklist on the file from grub4dos as requested?

 

re. drive C: -  do you have lots of spare space on drive C: and was it defragged recently?

[r4ymonf]

Oh, sorry about the blocklist part, I forgot about that.

The output was (hd0,0)7817512+7815832,536,4767081.

No, and no. :/

[steve6375]

The output was (hd0,0)7817512+7815832,536,4767081.

Are you sure that is accurate? There should be two + signs in the line????

e.g. 

[r4ymonf]

Oh, my mistake. It was (hd0,0)7817512+7815832,536+4767081.

[steve6375]

That is an odd result in that it lists the furthest-most sequential cluster run first. Was this a copy made using Windows Explorer or TeraCopy?

I really want to see just a plain Windows Explorer COPY blocklist.

Did you try the other experiments, i.e. repeat it and see if blocklist result is identical, try safe mode copy, etc.?