On Tue, Jun 07, 2011 at 12:15:55PM +0200, Jiri Denemark wrote:
> On Mon, Jun 06, 2011 at 14:03:44 -0300, Marcelo Tosatti wrote:
> ...
> > + "return":[
> > + {"device":"ide1-hd0",
> > + "status":"active",
> > + "info":{
> > + "percentage":23,
> > + }
> > + },
>
> What about using the same form of progress reporting as used by query-migrate?
> That is, instead of
>
> "info":{
> "percentage":23
> }
>
> the following would be reported:
>
> "disk":{
> "transferred":123,
> "remaining":123,
> "total":246
> }
>
> One can trivially compute percentage from that but it's impossible to get this
> kind of data when only percentage is reported. And total can even change in
> time if needed (just like it changes during migration).
Done.
> > + {"device":"ide1-hd1",
> > + "status":"failed"
> > + }
>
> Is there any way we can get the exact error which made it fail, such as EPERM
> or ENOSPC?
The errors that can generate "failed" here are internal to QEMU, so i
don't think exporting them to management application makes sense. This
are errors such as EIO from file system.
Error processing should be done in the block_copy command, that is where
checks of the destination image are performed.
