On Tue, Apr 16, 2013 at 12:22 AM, Eric Blake <eblake(a)redhat.com> wrote:
[...]
This indeed feels like a reasonable project, where there is enough
work
to do to fill up 12 weeks of good effort while still having something
measurable to commit, and where the design is known enough in advance
that it won't be stalled by developer discussion on correct design.
I will give my design plan for this project here which developers can
discuss my design so that we can get a good conclusion.
While the work should definitely start with offline renames, there
is
also room for some additional scope such as online rename support for
hypervisors such as ESX that might already provide such functionality
natively and where the existing libvirt approach of dumpxml/define may
be inadequate to expose the full hypervisor capability.
Yup. Online rename support is also included for this project. Osier has
rephrased the summary like following, let me add online rename support.
Project name: Add rename APIs for all objects
Summary: Add rename APIs for all objects, including domain, snaphot,
network, interface, nwfilter, secret, storage pool, storage volume and
online rename support for hypervisors, etc.
[...]
Remember that not all good ideas will be selected, but I think you
are
finally settling into a good idea. And even if you don't get accepted
by GSOC, your idea for contribution is still useful and would still be
worth working on.
Sure, actually, i love the storage portions of Libvirt very much. If not
accepted by GSOC, i will also finish my ideas for Libvirt. In fact, I wanna
do some deserved jobs(about storage knowledge) for Libvirt community.
Very thanks for your comments ;-)
--
Thanks
Harry Wei