Kuryr Kubernetes Dragonflow Integration

Kuryr Kubernetes Dragonflow Integration

Dragonflow is a distributed, modular and extendable SDN controller that enables to connect cloud network instances (VMs, Containers and Bare Metal servers) at scale.

Dragonflow adopts a distributed approach to mitigate the scaling issues for large scale deployments. With Dragonflow the load is distributed to the compute nodes running local controller. Dragonflow manages the network services for the OpenStack compute nodes by distributing network topology and policies to the compute nodes, where they are translated into Openflow rules and programmed into Open Vswitch pipeline. Network services are implemented as Applications in the local controller. OpenStack can use Dragonflow as its network provider through the Modular Layer-2 (ML2) Plugin.

Integrating with Dragonflow allows Kuryr to be used to bridge containers and VM networking in an OpenStack deployment. Kuryr acts as the container networking interface for Dragonflow.

Testing with DevStack

The next points describe how to test OpenStack with Dragonflow using DevStack. We will start by describing how to test the baremetal case on a single host, and then cover a nested environemnt where containers are created inside VMs.

Single Node Test Environment

  1. Create a test system.

It’s best to use a throwaway dev system for running DevStack. Your best bet is to use either Fedora 25 or the latest Ubuntu LTS (16.04, Xenial).

  1. Create the stack user.
$ git clone https://git.openstack.org/openstack-dev/devstack.git
$ sudo ./devstack/tools/create-stack-user.sh
  1. Switch to the stack user and clone DevStack and kuryr-kubernetes.
$ sudo su - stack
$ git clone https://git.openstack.org/openstack-dev/devstack.git
$ git clone https://git.openstack.org/openstack/kuryr-kubernetes.git
  1. Configure DevStack to use Dragonflow.

kuryr-kubernetes comes with a sample DevStack configuration file for Dragonflow you can start with. You may change some values for the various variables in that file, like password settings or what LBaaS service provider to use. Feel free to edit it if you’d like, but it should work as-is.

$ cd devstack
$ cp ../kuryr-kubernetes/devstack/local.conf.df.sample local.conf
  1. Run DevStack.

Expect it to take a while. It installs required packages, clones a bunch of git repos, and installs everything from these git repos.

$ ./stack.sh

Once DevStack completes successfully, you should see output that looks something like this:

This is your host IP address: 192.168.5.10
This is your host IPv6 address: ::1
Keystone is serving at http://192.168.5.10/identity/
The default users are: admin and demo
The password: pass
  1. Extra configurations.

Create NAT rule that will cause “external” traffic from your instances to get rewritten to your network controller’s ip address and sent out on the network:

$ sudo iptables -t nat -I POSTROUTING 1 -s 172.24.4.1/24 -j MASQUERADE

Inspect default Configuration

In order to check the default configuration, in term of networks, subnets, security groups and loadbalancers created upon a successful devstack stacking, you can check the Inspect default Configuration.

Testing Network Connectivity

Once the environment is ready, we can test that network connectivity works among pods. To do that check out Testing Network Connectivity.

Nested Containers Test Environment (VLAN)

Another deployment option is the nested-vlan where containers are created inside OpenStack VMs by using the Trunk ports support. Thus, first we need to deploy an undercloud devstack environment with the needed components to create VMs (e.g., Glance, Nova, Neutron, Keystone, …), as well as the needed Dragonflow configurations such as enabling the trunk support that will be needed for the VM. And then install the overcloud deployment inside the VM with the kuryr components.

Undercloud deployment

The steps to deploy the undercloud environment are the same as described above for the Single Node Test Environment with the different sample local.conf to use (step 4), in this case:

$ cd devstack
$ cp ../kuryr-kubernetes/devstack/local.conf.pod-in-vm.undercloud.df.sample local.conf

The main differences with the default dragonflow local.conf sample are that:

  • There is no need to enable the kuryr-kubernetes plugin as this will be installed inside the VM (overcloud).
  • There is no need to enable the kuryr related services as they will also be installed inside the VM: kuryr-kubernetes, kubelet, kubernetes-api, kubernetes-controller-manager, kubernetes-scheduler and kubelet.
  • Nova and Glance components need to be enabled to be able to create the VM where we will install the overcloud.
  • Dragonflow Trunk service plugin need to be enable to ensure Trunk ports support.

Once the undercloud deployment has finished, the next steps are related to creating the overcloud VM by using a parent port of a Trunk so that containers can be created inside with their own networks. To do that we follow the next steps detailed at Boot VM with a Trunk Port.

Overcloud deployment

Once the VM is up and running, we can start with the overcloud configuration. The steps to perform are the same as without Dragonflow integration, i.e., the same steps as for ML2/OVS:

  1. Log in into the VM:

    $ ssh -i id_rsa_demo centos@FLOATING_IP
    
  2. Deploy devstack following steps 3 and 4 detailed at How to try out nested-pods locally (VLAN + trunk).

Testing Nested Network Connectivity

Similarly to the baremetal testing, we can create a demo deployment at the overcloud VM, scale it to any number of pods and expose the service to check if the deployment was successful. To do that check out Testing Nested Network Connectivity.

Creative Commons Attribution 3.0 License

Except where otherwise noted, this document is licensed under Creative Commons Attribution 3.0 License. See all OpenStack Legal Documents.