Revision history [back]

click to hide/show revision 1
initial version

OH, By convention, you would create a bridge for internal traffic, for your tunnels, and for each physical network device you have attached. For example if I had 2 flat interfaces eth1 and eth2, and then an IP for my tunneling on eth3, then my OVS bridges would look like br-int br-tun br-eth1 br-eth2

While there is no set way to name and organize the bridges it does help to keep things clean. Also if you are using the ovs-agent, it will want to know what bridge you are using for internal and for tunnels, as well as physical networks for providers networks.

 grep -v "#" /etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini | grep -v '^$'
[ovs]
network_vlan_ranges = physnet1
enable_tunneling = True
integration_bridge = br-int
tunnel_bridge = br-tun
local_ip =172.17.8.48
bridge_mappings =physnet1:br-ex
[agent]
polling_interval = 2
tunnel_types =vxlan
vxlan_udp_port =4789
l2_population = False
arp_responder = False
enable_distributed_routing = False
[securitygroup]
firewall_driver = neutron.agent.linux.iptables_firewall.OVSHybridIptablesFirewallDriver

As you can see in the example above br-int and br-tun are defined by the plugin, and provider network physnet1 is mapped to br-ex.
All of these bridges need to be created using the #ovs-vsctl add-br <bridge> or you will get errors on the startup of the ovs agent.