Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

Different usecases using ML2

Hi guys,

We are setting up a 4 node setup (1 Controller node, 2 Compute Nodes (CN1, CN2), 1 Network Node) to view ML2 functionality on Havana. We are trying the following use cases to understand its functionality.

Note on Controller node configuration: We have set the following ML2 configuration:

           type_drivers = vlan, vxlan
           mechanism_drivers = openvswitch, linuxbridge

Use Cases:

  1. CN1 using OVS agent with VXLAN configured, CN2 using OVS agent with VLAN configured

  2. CN1 using linux-bridge agent with VXLAN configured, CN2 using linux-bridge agent with VLAN configured

  3. CN1 using OVS agent with VXLAN configured, CN2 using linux-bridge agent with VLAN configured

In this scenario, we would like to know the agents that needs to be run on Network Node (NN) – Is it OVS agent OR linuxbridge?

We have heard people talking that both the agents cannot run concurrently on the NN. So, how can we accomplish the above use cases if only one agent is allowed to run on the NN.

Is there any other configuration settings that needs to be done to enable it? Please let us know, how to move forward.

Also, please let us know is there any documentation or guide available for ML2 configuration on CentOS or RHEL , because it seems the packages for ML2 available in EPEL repo is not stable which make the neutron-server dead after configuring ML2. http://repos.fedorapeople.org/repos/openstack/openstack-havana/epel-6/

Alternatively , it would be great if we could get the debian packages for ubuntu as soon as possible.( ml2 missing) http://ubuntu-cloud.archive.canonical.com/ubuntu/dists/precise-updates/havana/main/binary-amd64/Packages

Thanks in Advance.

-- Regards, Romil