> I don't see any reason why it wouldn't work with VXLAN. Well with the current approach in your code you would have to basically re-implement all of the logic in the openvswitch agent that sets up VXLAN tunnels and isolation of nodes on different VXLAN networks (either via flows or local VLANs). That will be quite a bit of work.
We can already leverage the L2 agent’s logic if, instead of setting VLANs/VXLANs directly in your driver, we just wire up a patch port to br-int using the correct name that the L2 agent expects for the Neutron port the bare metal VM has. Then the L2 agent does its normal logic to get the traffic onto the right VLAN/VXLAN/GRE/whatever.
|
__________________________________________________________________________ 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