(Sorry, Rich, I managed to miss reply-all.) These VMDK files are difference files from a baseline VMDK file. I'm not familiar with ESX's storage, but the size indicators on these tell me they aren't raw. For a virtual disk that has 3.2GB of data in its current state, against 3.0GB of data in its baseline, the current state's .vmdk file is 200MB. --Alex On Mon, Jun 10, 2013 at 6:11 AM, Richard W.M. Jones <rjones@redhat.com>wrote:
On Sun, Jun 09, 2013 at 05:49:44PM -0400, Alexander Nelson wrote:
I've used QEMU to read .vmdk snapshots. The VM directory layout in my case (Fusion, and I've seen Workstation do the same) created a .vmdk file per snapshot, and qemu-img could use that .vmdk file and the base .vmdk to convert the disk image to a raw image. IIRC there is a manifest file that ties .vmdk files to the snapshot they represent.
So, from my experience, qemu does read disk snapshots in the .vmdk format. It might need a little extending to also look for and parse that manifest file, to treat a directory with .vmdk's as a snapshot tree.
Are these real VMDK files, or do you mean the "vmdk" files which are really raw that ESX produces?
TBH I don't know a lot about the VDMK format, except that the specifications are intimidating.
Rich.
-- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Fedora Windows cross-compiler. Compile Windows programs, test, and build Windows installers. Over 100 libraries supported. http://fedoraproject.org/wiki/MinGW