I have four 2gbsparse format VMDK files without the correct descriptor file. I need to access the data (part access will do). I do not know the correct original size of the whole disk or the last chunk.
The sizes of the chunks are:
2,094,596,096 test-c-export-s001.vmdk
2,139,947,008 test-c-export-s002.vmdk
1,790,012,344 test-c-export-s003.vmdk
1,739,020,288 test-c-export-s004.vmdk
Can any tool cope with this situation? I have not managed to persuade VDK to read the chunks yet.
the descriptor you made up is inconsistent - you use a diskgeometry for a 38 Gb disk.
You should better use
ddb.geometry.cylinders = "1043"
instead of 5000.
Do not use vmkfstools - i would mount it with vdk.exe instead.
Anyway - I doubt that it will work like this - the disk probably was not created in the size of 8188 MB.
I suggest to fake a 5th slice and use a geometry for 8 Gb
___________________________________
VMX-parameters- WS FAQ -[ MOAcd|http://sanbarrow.com/moa241.html] - VMDK-Handbook
Is that all you have ? - if you have old vmware.logs see my site http://sanbarrow.com/vmdk-howtos.html#restorefromlog
Are you sure that you got all the slices ? the problem is guessing the correct nominal size - for the 4 slices this is probably slightly less then 8 GB.
But from the looks I would guess you do not have all slices ?
tricky - there exists no tool that can help you here ...
I could guess one for you but that may need several attempts ...
Suggestion: lets try to reconstruct a 8 Gb disk - then I need to know which filesystem you used inside the vmdk ? - NTFS ?
Then I can fake the 5th slice
___________________________________
VMX-parameters- WS FAQ -[ MOAcd|http://sanbarrow.com/moa241.html] - VMDK-Handbook
I am not sure the below KB works or not?
Create a header for a 2gbsparse virtual disk
binoche, VMware VCP, Cisco CCNA
Hmm - that kb only explains the easy 5 % of the job
___________________________________
VMX-parameters- WS FAQ -[ MOAcd|http://sanbarrow.com/moa241.html] - VMDK-Handbook
Binoche, thank you this is helpful, although I think that the KB assumes you know some things like size of the last chunk that unfortunately I don't know.
Daniel
Thank you Continuum for your reply. I had already looked hard at your excellent site at . This file was being kept as a backup and therefore the logs are long gone. The disk is NTFS.
I think that it is an 8 GB file and I have made some progress. I have faked up an 8 GB descriptor file (below). The commented out lines are because VDK does not like them.
When I try to expand this disk using vmkfstools -i it counts up to 71% twice and then fails with the message
Clone: 71% done.Failed to clone disk : Bad file descriptor (589833).
I am wondering if one of the chunks is corrupt or if I have the number of bytes correct for the last chunk (4192256).
If I manually stop the vmkfstools process just before it fails I am left with a VMDK file that seems to have the directory in a perfect state but no actual data. This is some form of progress, I guess.
<CODE>
#Disk DescriptorFile
version=1
CID=676bdb76
parentCID=ffffffff
createType="twoGbMaxExtentSparse"
#Extent description
RW 4192256 SPARSE "test-c-export-s001.vmdk"
RW 4192256 SPARSE "test-c-export-s002.vmdk"
RW 4192256 SPARSE "test-c-export-s003.vmdk"
RW 4192256 SPARSE "test-c-export-s004.vmdk"
#The Disk Data Base
#DDB
#ddb.deletable = "true"
ddb.virtualHWVersion = "4"
ddb.toolsVersion = "7302"
#ddb.encoding = "UTF-8"
#ddb.longContentID = "4805e269452961f9f0a28bb2383f79ba"
#ddb.uuid = "60 00 C2 9f 15 1b c6 68-59 d6 29 7f 83 23 9e fd"
ddb.geometry.cylinders = "5000"
ddb.geometry.heads = "255"
ddb.geometry.sectors = "63"
ddb.adapterType = "lsilogic"
</CODE>
the descriptor you made up is inconsistent - you use a diskgeometry for a 38 Gb disk.
You should better use
ddb.geometry.cylinders = "1043"
instead of 5000.
Do not use vmkfstools - i would mount it with vdk.exe instead.
Anyway - I doubt that it will work like this - the disk probably was not created in the size of 8188 MB.
I suggest to fake a 5th slice and use a geometry for 8 Gb
___________________________________
VMX-parameters- WS FAQ -[ MOAcd|http://sanbarrow.com/moa241.html] - VMDK-Handbook
Using 1043 cylinders I can now read the disk using VDK. This is a total success. Yesterday I was using 1044 which appeared to be correct from the geometry tables, but I must have read it wrong.
Thanks very much Continuum!
Daniel
1044 would be ok for a 8 Gb disk - but your 4 slices are only 4 x 2047 MB = 8188
you were lucky - the missing 4 Mb fit into the unformatted space at the end of the disk so your partition-table was still appropriate
___________________________________
VMX-parameters- WS FAQ -[ MOAcd|http://sanbarrow.com/moa241.html] - VMDK-Handbook
cool continuum!
binoche, VMware VCP, Cisco CCNA