Open Stack

Hi,

Yes you should setup a trunk port on your physical switch for your compute and network nodes.
On network and computes nodes openvswitch will do the trick if well configured.

Here is  a sample configuration of your /etc/neutron/plugin.ini
network~ # ls -l /etc/neutron/plugin.ini 
lrwxrwxrwx. 1 neutron neutron 42 Feb 10 14:32 /etc/neutron/plugin.ini -> plugins/openvswitch/ovs_neutron_plugin.ini
network~ #

[OVS]
integration_bridge = br-int
tenant_network_type = vlan
enable_tunneling = False
network_vlan_ranges = physnet1:111:222
#bridge_mappings = physnet1:br-eth0
bridge_mappings = physnet1:br-eth1

which means that eth2 is connected to a trunk port on the physical switch.
 

> -----Original Message-----
> From: Erich Weiler [mailto:weiler at soe.ucsc.edu]
> Sent: lundi 7 avril 2014 09:47
> To: openstack
> Subject: [Openstack] VLAN configs
> 
> Hey Y'all,
> 
> If I'm using VLAN as network_type for my internal tenant networks, I assume
> I need to configure the switchports on my switch for the compute nodes and
> network node as trunk ports yes?  Do I need to do any special trunk config to
> my data interfaces on my compute nodes and network nodes?  Or does
> Open vSwitch handle that for me on the compute node side?  The online
> guides don't mention anything about that, so I'm asking....
> 
> I used the havana guide for redhat (VLANs and Open vSwitch) and I get no
> network access (inside the local network, can't even ping gateway), so I'm
> assuming there is some kind of VLAN issue.  VMs also don't get DHCP
> addresses, presumably because they have no local network access for DHCP.
> 
> Thanks,
> erich
> 
> _______________________________________________
> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
> Post to     : openstack at lists.openstack.org
> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack