[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Announcing XenMaster

On Mon, Feb 13, 2012 at 10:14 AM, Grant McWilliams
<grantmasterflash@xxxxxxxxx> wrote:
>> >  It totally makes sense
>> > for tools that aim to manage multiple types of hypervisors or multiple
>> > types of storage.
>> > For a XCP frontend it makes not much sense to base it libvirt since
>> > the direct XAPI access is more suited to manage SRs and other
>> > specialties and XCP does already abstract all of those.
>> Agreed. There are a lot of features that a generic tool like libvirt
>> simply doesn't support (as it is a common abstraction layer) and not
>> purpose-built for Xen like XAPI and libxl are.
>> Here is a summary of the choice of toolstacks:
>> http://wiki.xen.org/wiki/Choice_of_Toolstacks
> It looks like Todd that in the future XAPI and libvirt will both be using
> libxl underneath?

libvirt already has a libxl port. XAPI's libxl port is still in progress.

> I would also assume that this won't change how XAPI works
> for a user/administrator but rather how it communicates with the Hypervisor?

Right, the point of libxl is doing the "bottom third" of any Xen-based
toolstack and doing it well. XAPI, libvirt, or any other toolstack can
then innovate on top of that. Porting XAPI to a libxl base will likely
just make it faster, more maintainable, and compatible with
libxl-based development. The admin or developer API will not be
affected by the libxl port.


Todd Deshane

Xen-users mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.