I'm running Fusion 2.0.5 on a 2.8Ghz Intel Core Duo iMac running OS 10.5.8. The operating system on the VM is XP Pro SP3. I'm trying to minimize the "memory footprint" of the VM and have "moved" the MyDocuments folder to point to a folder that is native on the MacOS (shared folder through Fusion). The MyDocuments contents total about 414MB. The only files resident on the VM are the operating system and the XP programs. All my user data is stored on the MacOS thru a shared folder in fusion. When I originally set up the VM I set it as a 40GB machine. At the time that I installed fusion and set up the VM I recall reading that, although the VM was set up at 40GB, the software would only create source files to "house" the actual data in the VM - meaning it wouldn't take up 40GB unless there was 40GB worth of data in the VM. I do not have any snapshots. With that background, here's the dilemma:
When I start the VM and "add up" the folder sizes through Windows Explorer, my data adds up to about 6.5GB. When I do a "properties" on the C: drive it indicates that there is 8.96GB of data on the drive. When I check the size of the VM on the MacOS it is a whopping 35.43GB and it appears to grow each time I launch and close the VM.
What kind of file system architecture is at play here?
Why are the VM source files 4 times the size of the "data" on the C: drive?
Why is the data indicated on the VM C: drive almost 2.5GB larger than the actual data in the VM?
How do I "shrink" the source files to accurately represent the quantity of the actual data in the VM?
I'm trying to optimize my backup scheme and want to make sure the size of the VM is minimized at all times.
Yes, I know about 'sparse bundles' and the fact that the new version allows you to split the source files into 2GB chunks. That's not the point of this post.
I want to know why the source files are several orders of magnitude larger than the actual data in the VM and how to shrink them back to accurately reflect the data in the VM. I would also like to keep the source files from self inflating - which they appear to do even when no data has been added to the VM. Any insight into this would be greatly appreciated.