于 2011年08月17日 22:59, Daniel P. Berrange 写道:
On Wed, Aug 17, 2011 at 09:10:41PM +0800, Osier Yang wrote:
> If one tries to restore a domain from a corrupt save image, we blindly
> goes forward to restore from it, this can cause many different errors,
> depending on how much the image is saved. E.g.
>
>
https://bugzilla.redhat.com/show_bug.cgi?id=730750
>
> So I'm thinking if we can introduce a new feild to struct qemud_save_header,
> such as "bool complete;", and set it true if succeeded to save the image,
> false if not. So that could do some checking while trying to open the image
> (qemuDomainSaveImageOpen), and quit early if "complete" is false, with
> a sensiable error message.
>
> Thought?
I assume you mean that when saving the guest, we'd do the following sequence
1. Write out basic header with complete=false
2. Write out XML doc
3. Run 'migrate' in QEMU to write save state
4. If success, update header with complete=true
And then on restore we'd do
1. If complete == false, then quit with error
2. Run QEMU with -incoming to restore
At which point I wonder what's wrong with:
1. Write out basic header
2. Write out XML doc
3. Run 'migrate' in QEMU to write save state
4. If not success, unlink save fail
The only case where there's any difference, is if libvirtd itself crashes
between steps 1& 4.
Daniel