[one-users] Backing up SSH system datastore disk images

[Mr Sensible]

I currently run a couple of hosts with an SSH system datastore (and 
remote image store), as there are some VMs with intensive disk activity, 
and I don't want to kill my network (no fancy fibre SAN for me :( ).

I understand that once I deploy the VM, the vm disk image is copied to 
the host's system datastore, and I have set it to persistent, so state 
should carry over to VM restarts (I think, in the process of testing).

As such, I understand that there is no live migrate to recover from host 
failure, but I would be reluctant to lose state of the vms in the local 
system datastore.

I have tried to read the various documents, but it is unclear to me what 
state a VM has to be set to to copy the vm image back from the System 
datastore to the remote Image datastore. Ideally I think I need to kill 
the VM at midnight in such a way that it will copy the image back to the 
remote Image datastore, and then "restart/deploy/thing" the VM so it is 
available during work time.

I may be overcomplicating things here, but a couple of questions:
- What state should I set the VM to flush the disk image back to the 
remote image datastore?
- Is there a simpler method of flushing the disk image back to the 
remote image datastore?

Thanks in advance

Peter