The Server is a fully patched 2003 Server. P2V should be working, everything is full 1GB, all drives are NTFS, nothing funny. I'm doing this all hot by the way, no cold booting, its a production server with no downtime allowed. Each P2V attempt died near the end, it takes several hours, so I was never watching exactly when it died. Checking the logs on the client at C:\Documents and Settings\All Users\Application Data\VMware\VMware Converter Enterprise\Logs\vmware-converter-agent-2.log . Shows me the error :
"Image processing task has failed with PlatformError fault: (1117) The request could not be performed because of an I/O device error.".
Obviously this indicates an I/O error. After some searching I found an interesting fact, if you maintain the size of the original drive the P2V or "import machine" will do a block level copy, and if there are bad blocks, then you can get this error. So if you change the destination drive size, it does a file level copy instead. There is a chance that this server's hard drive has bad blocks, but honestly I don't know. It's an old server, that's why i'm P2Ving it. After making a change to the destination partition size, this fixed my issue and the P2V was finally sucessful.
FYI, this was all done with the latest software, ESX 4.0U1, VC 4U1, etc...