Try this, I found it very useful. https://openstack.redhat.com/Networking_in_too_much_detail
Danny Date: Sun, 14 Dec 2014 21:25:24 +0100 From: Uwe Sauter <uwe.sauter...@gmail.com<mailto:uwe.sauter...@gmail.com>> To: Gonzalo Aguilar Delgado <gagui...@aguilardelgado.com<mailto:gagui...@aguilardelgado.com>> Cc: openstack@lists.openstack.org<mailto:openstack@lists.openstack.org> Subject: Re: [Openstack] What's physical network for? Message-ID: <548df234.4060...@gmail.com<mailto:548df234.4060...@gmail.com>> Content-Type: text/plain; charset=utf-8 As far as I know, bridge_mapping is used to let neutron know which bridge should be uses for the (possibly more than one) external network(s). If someone else could jump in and give a bit more detail, I'd appreciate that, too. Regards, Uwe Am 14.12.2014 um 20:55 schrieb Gonzalo Aguilar Delgado: Hi Uwe, really I started directly with neutron. Never gone with legacy. But I suppose there's old config laying around. I still think that bridge_mapping is needed for VLAN config I use. Every paper describes GRE config, but I still feel more confortable using VLANs. Any other help about this issue? Can someone confirm if I can get rid of the directives in both configs? I suppose I cannot because they get in effect when started the ovs plugin. Thank you in advance. El dom, 14 de dic 2014 a las 8:40 , Uwe Sauter <uwe.sauter...@gmail.com<mailto:uwe.sauter...@gmail.com>> escribi?: Hi, I presume that you upgraded from an older version that used nova-network (now called legacy networking). Using neutron means that VMs aren't connected to br0 directly any more as there is a whole virtual networking infrastructure in place. To give a small overview: On a compute node a VM connects to br-int (integration bridge). This bridge itself is connected through a virtual cable to br-tun (tunneling bridge). That bridge has also assigned a physical interface that allows traffic to flow to the network node On the network node there also exists a br-tun that has a physical interface attached. Through this inferface traffic enters the node. br-tun is virtually connected to br-ex that has a separate physical interface attached that connects to "the outside", meaning the networking infrastructure outside your cloud. I cannot help you with the configuration issue but recommend that you familiarize yourself with neutron. Regards, Uwe Am 14.12.2014 um 19:36 schrieb Gonzalo Aguilar Delgado: Hi all, I'm installing a new compute node from scratch and reviewing all old config. I've found two setting that seems equal, one in ml2 plugin and one in openvswitch. But I don't really understand why they are. ovs_neutron_plugin.ini: bridge_mappings = default:br0,extnet1:br-ex ml2/ml2_conf.ini: [ovs] bridge_mappings = default:br0,extnet1:br-ex For me it's strange the settings are in both places. I think this is a result of upgrading without taking much care of removing old config. But also it's strange that everything works with the bridges br0 and br-ex without physical interface. I mean, seems to do nothing but it needs to be there. Also I should expect VM be attached to br0 (Default) but it's not, they are attached to the br-int (integration bridge), for me this is correct. Since it's described here like this: https://openstack.redhat.com/Networking_in_too_much_detail And works ok. So what's the purporse of these bridges? Here is: neutron 2.3.4 nova 2.17.0 Best regards, _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org<mailto:openstack@lists.openstack.org> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org<mailto:openstack@lists.openstack.org> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
_______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
