Why does kolla deployment need two NICs?

asked 2017-09-28 15:06:08 -0600

jcorkey gravatar image

I can't find much documentation explaining this and I am very new to using kolla. The quick start guide for setting up a deployment says each node needs two NICs.

I have three servers, a deploy host, a controller node, and a compute node. I am using kolla-ansible to do a simple multinode deployment on the controller node and compute node and I am wondering why each node needs two NICs?

What exactly does each NIC get used for during the deployment?

edit retag flag offensive close merge delete


I believe the production architecture guide do tell you how each nic is used https://docs.openstack.org/kolla-ansi...

iGene gravatar imageiGene ( 2017-09-28 21:18:57 -0600 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2017-09-29 03:09:33 -0600


First NIC is for management (network_interface) most interfaces defaults to this interface (api_interface,tunnel_interface,etc). This is an interface used for API request, user access, VIP creation and have an IP associated with it.

Second interface is for neutron external network (neutron_external_interface), this interface will be used by OpenvSwitch and will be attached to br-ex.

If both interfaces are the same, when the interface is attached to br-ex, will loose all communication with the host, as the IP address on it will be removed because is fully managed by ovs and not the host.

For the nodes who need the second interface, it depends of what your network architecture will be. If will use default tenant network + public, only network nodes will need neutron_external_interface, while if using providers networks also compute nodes need it.

In you case I guess control also contains network, so this node will need the second interface when using tenant networks.


edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2017-09-28 15:06:08 -0600

Seen: 679 times

Last updated: Sep 29 '17