On Wed, Mar 28, 2007 at 11:10:38AM -0400, Daniel Veillard wrote:
On Wed, Mar 28, 2007 at 03:55:24PM +0100, Daniel P. Berrange wrote:
> On Wed, Mar 28, 2007 at 10:45:34AM -0400, Daniel Veillard wrote:
> > It would be nice though if each time you look whether the domain is
> > defined or not, libvirt could give teh answer directly instead of going
> > though the potentially expensive Info call. This would also require
> > to double check invalidation of the _virDomain cached data at the
> > proper places.
>
> The virDomainGetInfo is a fast-path call - its either a direct hypercall,
> or a hypercall via the proxy - we never make an expensive XenD call unless
> they've broken the proxy somehow.
I was thinking about the upcoming remote support, and an RPC is expensive
for this kind of informations IMHO.
The only way to avoid that overhead then would be to add an extra bit of
info to the virDomainPtr struct, which is populated at the time the virDomainPtr
object is allocated, so we can avoid calling back into the driver model.
That would raise some troublesome cache invalidation issues on the client
end, though I guess we've already got that issue with the ID value.
Dan.
--
|=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=|
|=- Perl modules:
http://search.cpan.org/~danberr/ -=|
|=- Projects:
http://freshmeat.net/~danielpb/ -=|
|=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|