On Jul 2, 2013, at 8:26 AM, Alex Gaynor <[email protected]>
wrote:
> I think I've been swayed to favor (1) (putting it on compute drivers). a)
> it limits the code churn, which is nice, b) it's obvious they need *some*
> integration with each other, and there aren't well established patterns for
> that right now, rather than invent something new, grouping them together
> makes a lot of sense.
The only way to attach a block storage volume to an instance in
OpenStack is via the compute API (at least for now), so integrating with
compute makes a lot of sense.
-- Ed Leafe
