I have tried (and rolled back) a P2V of a win2003r2 server. Everything seemed fine with IDE virtual disks, but when backup task (snapshot) started on the VM, I got a lot of IDE timeouts in the windows guest log, then the VM crashed. I found the VM simply switched off. Restarted it, found the errors in the log, switched to SCSI virtual disk, restarted, tried to run a backup again, and again I got timeouts in windows log. This time I also got ntfs corruption (indicated in the logs, and verified by a checkdisk).
At this point I rolled back to the phisical, then I have run more tests on the abandoned virtual installation. I have found that using virtio (not scsi) disk seems to avoid corruption, even if I still get event 129 in windows logs (which I googled and it means timeout) while running a snapshot backup.
Backups are made to a NFS remote storage via a gigabit network, and the nfs server is capable of sustained full gigabit tranfer rate.
Anyway the issue is with the timeouts and filesystem corruption. This of course is an enormous issue.
At this point I rolled back to the phisical, then I have run more tests on the abandoned virtual installation. I have found that using virtio (not scsi) disk seems to avoid corruption, even if I still get event 129 in windows logs (which I googled and it means timeout) while running a snapshot backup.
Backups are made to a NFS remote storage via a gigabit network, and the nfs server is capable of sustained full gigabit tranfer rate.
Anyway the issue is with the timeouts and filesystem corruption. This of course is an enormous issue.