Il 10/02/2015 16:49, Sven Kieske ha scritto:
On 10/02/15 03:02, Jason Brooks wrote:
The meaning of "support" is important here -- support from whom? It's true that there's no gluster+virt SKU of the RHEV downstream project. All configurations of ovirt proper are self-supported, or community- supported, and what we choose to support is up to us individuals.
However, gluster + virt on the same nodes does work -- even w/ management through the engine. I do use gluster on my virt nodes, but I don't manage them w/ the engine, because, afaik, there isn't a way to have gluster and virt on separate networks this way, so I just manage gluster from the gluster cli.
It's true, oVirt is happiest w/ separate machines for everything, and a rock-solid san of some sort, etc., but that's not the only way, and as you point out, hardware isn't free.
Well you might be interested in these upcoming features:
http://www.ovirt.org/Features/Self_Hosted_Engine_Hyper_Converged_Gluster_Sup...
sadly the slides from this talk are not online: https://fosdem.org/2015/schedule/event/hyperconvergence/
maybe brian(cc'ed) can put them somewhere?
Or Federico (CCed) :-)
-- Sandro Bonazzola Better technology. Faster innovation. Powered by community collaboration. See how it works at redhat.com