On 07/30/10 - 10:43:01AM, Luiz Capitulino wrote: > On Fri, 30 Jul 2010 11:34:57 +0200 > Markus Armbruster <arm...@redhat.com> wrote: > > > Miguel Di Ciurcio Filho <miguel.fi...@gmail.com> writes: > > > > > This patch address two issues. > > > > > > 1) When savevm is run using an previously saved snapshot id or name, it > > > will > > > delete the original and create a new one, using the same id and name and > > > not > > > prompting the user of what just happened. > > > > > > This behaviour is not good, IMHO. > > > > Debatable. > > Automatically destroying previously saved data without any notice seems > a quite bad behavior to me. > > > > We add a '-f' parameter to savevm, to really force that to happen, in > > > case the > > > user really wants to. > > > > Incompatible change, looks like it'll break libvirt. Doesn't mean we > > can't do it ever, but right now may not be the best time. Perhaps after > > savevm & friends are fully functional in QMP. > > Chris, could you please check whether this impacts libvirt?
Sorry for the delay here. As far as libvirt is concerned, this won't break the functionality, just change the semantics. If you only ever do snapshots through libvirt, then we can never run into this situation; libvirt prevents 2 snapshots from having the same name. Today, if you do a mixture of snapshots through the monitor and snapshots through libvirt, and you name them the same, then libvirt *could* silently override the old one. After this patch, the savevm will fail (which libvirt will gracefully handle. In any case, it's a corner case that libvirt will never intentionally put itself into, so either way is fine with me. (I also tend to think that overwriting data without any notification isn't very nice, but I also understand that this is a change in semantics) -- Chris Lalancette
