Stephen, Michael, thank you for having a look. I'll respond to every issue you mentioned when I get to work on Sunday.
Until then, in case you don't mind inspecting a small diff, just to clarify my point, please have a look at a rather straightforward change, which 1. exemplifies pretty much all I'm currently proposing (just splitting amphorae into semantic sub-clusters to facilitate code-reuse) 2. I'm hoping should provide everything needed (and thus frictionless review) for the virtual non-shared distributor of active active topology 3. is quite transparent for other topologies, including future active-active shared, hardware, what-have-you, just because it's fully compliant with existing code https://github.com/sgserg/octavia/commit/030e786ce4966bbf24e73c00364f167596aef004 Needless to say, I wouldn't expect anything like this to be merged until we see an end-to-end working (virtual-private-d'tor) AA N+1 create-lb proof of concept (not destroying existing topologies). I'm not married to this idea, it's just something I came up with having spent a few weeks in front of the code, trying to imagine how the simplest active-active use-case would go around performing the same tasks (vrrp, vip plugging, etc.). -Sergey.
__________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev