This message is a reminder that Fedora 12 is nearing its end of life.Īpproximately 30 (thirty) days from now Fedora will stop maintainingĪnd issuing updates for Fedora 12. I think therefore the only problem is when the guest doesn't have ACPI enabled. Note that I created a new Windows XP VM (from the same install media) WITH ACPI enabled in the guest, and the shutdwon command and button function exactly as they should do. > Gleb, is it expected that 'system_powerdown' does nothing with XP guests, even > - on F-12, there is, but 'virsh shutdown' does nothing > - on F-11, there's no so 'virsh shutdown' would cause the guest > AFAICS, the behaviour with an XP guest is: > libvirt merely sends a 'system_powerdown' monitor command to qemu when you do So I tried commenting out the line, but there was no change in behaviour from what I describe above wrt shutdown, surprisingly. > Does your guest have in its section? Does adding it help? > In F-12, we enabled ACPI for XP guests: I'll re-open it for the time being so others can look at it, but I'm not at all certain it's something we can/should fix. You could argue that once the guest is shutdown (without ACPI), the only reasonable thing to do is to physically poweroff the machine, which is essentially what "virsh destroy" does. It's also possible that this was a deliberate change in behavior. It's possible that qemu inadvertently changed it's behavior, though. As far as I know, virsh shutdown has always just injected an ACPI event, so there wouldn't be much change there. It does seem like a change in behavior, although I can't imagine where. This worked in F-11, doesn't work inĪh, this is a different story. > equivalent of powering off a real machine. > the shutdown button within virt-manager) would stop the guest process - the > safe to shutdown your computer screen". > In F-11 I shutdown the guest from within the guest until I get the "It is now > Actually I think you may have missed my point, perhaps I should be clearer: > button worked without ACPI with the F-11 virt stack, so this is a regression. > Right - I understand what you're saying, but the fact is that the shutdown > I'm going to close this as NOTABUG, since there's nothing that libvirt (or even This worked in F-11, doesn't work in F-12. At that point shutdown within virsh (or the shutdown button within virt-manager) would stop the guest process - the equivalent of powering off a real machine. In F-11 I shutdown the guest from within the guest until I get the "It is now safe to shutdown your computer screen". Actually I think you may have missed my point, perhaps I should be clearer: Right - I understand what you're saying, but the fact is that the shutdown button worked without ACPI with the F-11 virt stack, so this is a regression. > change the guest, you may as well use the standard mechanisms. Either one is a change to the guest, and if you are going to > shutdown inside the guest, you'd need to install the paravirt driver in the While you could certainly have a paravirt driver that induced a > Note that this isn't directly solvable via some sort of paravirt drivers, > to have ACPI enabled and working inside the guest. > an ACPI button press event into the guest. The way that shutdown works for a kvm guest is to inject > I havea Windows XP guest instance, without ACPI support enabled in the guest Version-Release number of selected component (if applicable): In the end the only way to shutdown the guest is to kill the qemu process. The shutdown button in virt-manager and the shutdown command in virsh fail to shutdown the guest (in fact they don't seem to do anything). I havea Windows XP guest instance, without ACPI support enabled in the guest (though I don't think that is relevant).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |