I wrote:
Quoting Michael Paoli (Michael.Paoli@cal.berkeley.edu):
Yup ... one (or a few) step(s) closer to being able to better monitor such a potential boo-boo.
Just a sudden wild idea, not necessarily a practical one: Just as software tools in the _conventional_ (non-VM) use case to ensure that zonefile S/Ns always go forwards and never backwards have advantages, one might imagine a wrapper for VM image files that builds a timestamp into the filename and enforces always moving forwards (or not moving) in time and never going backwards.
Somehow, my concluding paragraph got accidentally deleted before I sent that. I said something like:
However, I strongly doubt that the problem actually merits any such programmatic guard measures, because this really seems like a problem unlikely to recur.
(I may have had an additional paragraph beyond that, too, but can no longer recall particulars.)