On Thu, May 16, 2013 at 11:58 AM, Klaus Aehlig <[email protected]> wrote:
> > So far, for computing reboot groups, hroller only considered > drbd-instances. Even though the non-drbd instances can be > migrated to any node, we have to account for capacity. As > suggested in the design document, this is done by greedily > searching for locations where to place the non-drbd instances. > > Is it? - HRoller should handle non redundant instances: currently these are ignored but there should be a way to select its behavior between "it's ok to reboot a node when a non-redundant instance is on it" or "skip nodes with non-redundant instances". This will only be selectable globally, and not per instance. Then in future work we talk about moving them. It's perfectly ok (and thanks) for moving them, but there should be an option to say "don't do those nodes" or "it's ok, just shut them down" as well. :) Thanks, Guido
