On Fri, Oct 02, 2009 at 02:06:12PM +0200, Paolo Bonzini wrote:
On 10/02/2009 01:53 PM, Daniel P. Berrange wrote:
>This bit of code is exactly why we need a new API - the return value is
>declared to be a virDomainPtr on the destination host, not just a copy
>of the original source domain which no longer even exists on the source
>host if it was a transient domain.
Yeah, I really did not like this.
In Chris's code he actually opens the connection (after the migration is
complete) in order to return the destination domain.
Does that make sense?
Not really, because one of the requirements from libvirt-qpid is that
we have an API that only needs a localhost connection for migration,
to avoid the need to configure client app security credentials.
This implies we cannot either require a destination virConnectPtr in
the API, not open one internally.
Daniel
--
|: Red Hat, Engineering, London -o-
http://people.redhat.com/berrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org -o-
http://ovirt.org :|
|:
http://autobuild.org -o-
http://search.cpan.org/~danberr/ :|
|: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|