Error when converting VmWare virtual disk to HyperV

Question

I received VmWare image from my client for some testing purposes. I need to convert it to Hyper-V. I followed steps in http://www.askme4tech.com/how-convert-vmware-virtual-machine-hyper-v. I installed Microsoft Virtual Machine Converter and started to convert virtual disks in PowerShell. However I'm getting eror:

C:\Windows\system32> ConvertTo-MvmcVirtualHardDisk -SourceLiteralPath "c:\temp\disk2.vmdk" -DestinationLiteralPath "c:\data\HyperV\PH\" -VhdType DynamicHardDisk -VhdFormat Vhdx

ConvertTo-MvmcVirtualHardDisk : The entry 1 is not a supported disk database entry for the descriptor.
At line:1 char:1
+ ConvertTo-MvmcVirtualHardDisk -SourceLiteralPath "c:\temp\disk2.vmdk"
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : WriteError: (Microsoft.Accel...nversionService:DriveConversionService) [ConvertTo-MvmcVirtualHardDisk], VmdkDescriptorParseException
    + FullyQualifiedErrorId : DiskConversion,Microsoft.Accelerators.Mvmc.Cmdlet.Commands.ConvertToMvmcVirtualHardDiskCommand

ConvertTo-MvmcVirtualHardDisk : One or more errors occurred.
At line:1 char:1
+ ConvertTo-MvmcVirtualHardDisk -SourceLiteralPath "c:\temp\disk2.vmdk"
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : WriteError: (Microsoft.Accel...nversionService:DriveConversionService) [ConvertTo-MvmcVirtualHardDisk], AggregateException
    + FullyQualifiedErrorId : DiskConversion,Microsoft.Accelerators.Mvmc.Cmdlet.Commands.ConvertToMvmcVirtualHardDiskCommand

Answer 1

Modern VMDK files have a main descriptor file and separate data files. There is no need to use dsfo/dsfi tools.

• Ubuntu.vmdk <- the descriptor

• Ubuntu-s001.vmdk, Ubuntu-s002.vmdk etc. <- data files

In this case You simply needed to comment out the line:

#ddb.toolsInstallType = "4"

Answer 2

I found some adhoc solution - a bit hack perhaps but it works at least.

Digging into similar issues found on google I came to a tool to extract disk descriptor out of the VMDK file. The content of the descriptor for my VMDK was something like this:

# Disk DescriptorFile
version=1
encoding="windows-1252"
CID=5379bf0f
parentCID=ffffffff
isNativeSnapshot="no"
createType="monolithicSparse"

# Extent description
RW 209715200 SPARSE "00054_C8PHS1096_151216-disk2.vmdk"

# The Disk Data Base 
#DDB

ddb.adapterType = "lsilogic"
ddb.geometry.biosCylinders = "13054"
ddb.geometry.biosHeads = "255"
ddb.geometry.biosSectors = "63"
ddb.geometry.cylinders = "13054"
ddb.geometry.heads = "255"
ddb.geometry.sectors = "63"
ddb.longContentID = "64d4e008b7227bcce8aa54995379bf0f"
ddb.toolsInstallType = "1"
ddb.toolsVersion = "10241"
ddb.uuid = "60 00 C2 96 f7 70 f2 fd-b5 02 9e 46 6c df 00 2e"
ddb.virtualHWVersion = "10"

The error message together with the content of the extracted descriptor came to my attention, specifically the line:

ddb.toolsInstallType = "1"

as it contains the strange value of 1 from my error message. I edited the descriptor - just comment out that single line with # (hash mark), injected it back into VMDK and voila - the conversion works now.

Credits to this link https://communities.vmware.com/thread/343214?start=0&tstart=0 and of course to tools by Dariusz Stanislawek.

Just for reference, the steps I have done:

• download and extract dsfok tools
• use dsfo.exe "c:\temp\disk2.vmdk" 512 1024 descriptor1.txt to extract the descriptor
• edit the descriptor file in Notepad++: comment the above mentioned line (as I added the extra single character (#) I also deleted one NULL character from the end to keep the file size of 1024 bytes (not sure if this is needed).
• use dsfi.exe "c:\temp\disk2.vmdk" 512 1024 descriptor1.txt to inject the descriptor back into the VMDK
• repeat these steps for the other disk (my VM has two .vmdk files)
• reissue the ConvertTo-MvmcVirtualHardDisk command

REMARK

After creating VM in Hyper-V, the machine didn't boot, it remained in black screen with fast blinking cursor (so called black screen of death). I don't know if it was caused by the conversion or by the fact that original disks in VMWare had been SCSI while I attached them as IDE. To fix it, I attached DVD with image of Windows and booted from DVD. I ran the Rapair system, started the command line and ran

bootrec.exe /fixBoot

Finally, the VM boots and runs.. end of story.

Answer 3

Here is an answer using powershell/.net which doesn't require you to download anything extra. This worked for me. There is a little written about it here: https://badflyer.com/vmdx-to-vhd-conversion/

There may be something in the VMDK descriptor which the conversion cmdlet doesn't understand. In which case, just commenting it out should be good enough.

If you don't have notepad++ installed, you'll need to change which editor is opened.

Step 1:

# Open VM-ware disk, read 1024 bytes at position 512
$vmdkFileName = 'D:\VM\CNC Windows 7 Professional\Windows 7 Professional-cl1.vmdk'
$vmdkFileStream = [System.IO.File]::Open($vmdkFileName, [System.IO.FileMode]::Open, [System.IO.FileAccess]::ReadWrite)
$vmdkFileStream.Position = 512

$bytes = [byte[]]::new(1024);
$vmdkFileStream.Read($bytes, 0, 1024)

# Write to a temp file
$tempPath = [System.IO.Path]::Combine([System.IO.Path]::GetTempPath(), [System.IO.Path]::GetTempFileName())
$tempfile = [System.IO.File]::OpenWrite($tempPath)
$tempfile.Write($bytes, 0, 1024)
$tempfile.Dispose()

# Open the editor. Wait for exit doesn't always seem to work for npp...
# Use whichever edit you like, it needs to show text, and also helpful if it can show whitespace/control characters
$editor = Start-Process 'C:\Program Files\Notepad++\notepad++.exe' -ArgumentList $tempPath -PassThru -Wait
$editor.WaitForExit()

# TODO, change what is causing the problem in the opened file.

If the error message was complaining about a "1", then look for lines in the descriptor which have the value "1". For example: ddb.toolsInstallType = "1" You can comment it out like:

# ddb.toolsInstallType = "1"

Then save the editor and write the bytes back into the VMDK. Make sure when you save, you're using an editor which doesn't change the encoding of the data...

Step 2:

# Read back the temp file
$tempfile = [System.IO.File]::OpenRead($tempPath)
$tempfile.Read($bytes, 0, 1024);
$tempfile.Dispose()

# Write back to the vmdk
$vmdkFileStream.Position = 512
$vmdkFileStream.Write($bytes, 0, 1024)

# Cleanup
$vmdkFileStream.Dispose();
del $tempPath

Now try the conversion again.

Answer 4

I tried the above solution, but in my descriptor there wasn't toolsInstallType property so it didn't worked (with the same error).

After some research I found this tool: https://cloudbase.it/qemu-img-windows/

The command was like this:

qemu-img.exe convert c:\PATH_TO_FILE\disk.vmdk -O vhdx c:\PATH_TO_FILE\disk.vhdx -p

The documentation for this tool is here QEMU Documentation the -p flag allows you to see progress of the convert.

Also I found it easier to work with since it's a portable tool. In this specific case it does not require also to fix the bootloader, I just attached the vhdx to the new VM and it started successfully.

Answer 5

I have another solution to this problem that worked for me so thought I would add it to the mix for any future readers.

I downloaded the StarWindConverter tool from here (its free).

https://www.starwindsoftware.com/tmplink/starwindconverter.exe

It supports multiple destination formats and gives you a nice wizard to walkthrough. Once you have selected all your options you get a progress screen like below:

The process of converting an offline .vmdk files to .vhdx was simple.

Answer 6

A complementary for answers above if you're still stuck with conversion.

In my case qemu-img.exe convert ... method succeed, but result .VHD image failed to start in Hyper-V:

Virtual hard disk files must be uncompressed and unencrypted and must not be sparse.

Use a hex editor (edit block 0x100..0x200) or the dsfi.exe/dsfo.exe way described above to comment out these attributes in your .VMDK file:

ddb.uuid.image="... Some guid here ..."
ddb.uuid.parent="00000000-0000-0000-0000-000000000000"
ddb.uuid.modification="00000000-0000-0000-0000-000000000000"
ddb.uuid.parentmodification="00000000-0000-0000-0000-000000000000"
ddb.comment=""

Just replace ddb -> #db.

After that, MMVC conversion succeed for me without failure, output image works:

Import-Module "C:\Program Files\Microsoft Virtual Machine Converter\MvmcCmdlet.psd1"
ConvertTo-MVMCVirtualHardDisk -SourceLiteralPath .\win10-32-disk1.vmdk -DestinationLiteralPath . -VHDFormat Vhdx -VHDType DynamicHardDisk

Answer 7

I had a this problem too trying to convert a VMWare image to VHD. My solution is similar to eXavier's, but I was able to do it with just a text editor.

The reason I could do a simpler fix was that the vmdk I got from VMWare was a small text file that referred to a number of other files. It looked like this:

# Disk DescriptorFile
version=1
encoding="windows-1252"
CID=4bd4d907
parentCID=ffffffff
isNativeSnapshot="no"
createType="twoGbMaxExtentSparse"

# Extent description
RW 8323072 SPARSE "Windows Server 2012-s001.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s002.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s003.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s004.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s005.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s006.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s007.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s008.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s009.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s010.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s011.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s012.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s013.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s014.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s015.vmdk"
RW 983040 SPARSE "Windows Server 2012-s016.vmdk"

# The Disk Data Base 
#DDB

ddb.adapterType = "lsilogic"
ddb.geometry.cylinders = "7832"
ddb.geometry.heads = "255"
ddb.geometry.sectors = "63"
ddb.longContentID = "439d288830654baf53d1f9594bd4d907"
ddb.toolsInstallType = "1"
ddb.toolsVersion = "10240"
ddb.uuid = "60 00 C2 97 21 a0 4e af-fc 21 68 15 2f 12 7f 22"
ddb.virtualHWVersion = "12"

The fix was to remove the #Disk Database line and all those below it, using a text editor, so that my file looked like this:

# Disk DescriptorFile
version=1
encoding="windows-1252"
CID=4bd4d907
parentCID=ffffffff
isNativeSnapshot="no"
createType="twoGbMaxExtentSparse"

# Extent description
RW 8323072 SPARSE "Windows Server 2012-s001.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s002.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s003.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s004.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s005.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s006.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s007.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s008.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s009.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s010.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s011.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s012.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s013.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s014.vmdk"
RW 8323072 SPARSE "Windows Server 2012-s015.vmdk"
RW 983040 SPARSE "Windows Server 2012-s016.vmdk"

It then converted to a VHD with no problems.