Skip to main content

Installing the Kubernetes OpenStack Cloud Controller Manager

Objective​

After following this guide you will be able to create publicly available endpoints from your Kubernetes cluster.

Background​

The OpenStack cloud provider allows the Kubernetes cluster to provision, monitor, and remove resources necessary for operation of the cluster. The OpenStack cloud provider implements interfaces defined by the Kubernetes cloud controller manager. The cloud controller manager is a daemon that embeds the cloud-specific control loops shipped with Kubernetes in the core. These loops were originally in the kube-controller-manager. However, cloud providers move at a different pace and schedule compared to the Kubernetes project, and abstracting the provider-specific code to the cloud controller manager allows cloud vendors to evolve independently from the core Kubernetes code.

After you create a Kubernetes cluster on OpenStack, you can use the OpenStack cloud provider to create a load balancer for your Kubernetes services. You can also use the cloud provider to create persistent volumes (PVs) backed by Cinder block storage and dynamically provisioned volumes.

For more information, please see the cloud provider documentation: OpenStack Cloud Controller Manager

Note: This documentation was written for Kubernetes version 1.24. If installing on a different Kubernetes version, please see the cooresponding documentation for that version: Cloud Provider OpenStack

Prerequisites​

  • A Kubernetes cluster on OpenStack

Generate application credentials​

First, you'll need to create an application credential for Kubernetes to use. Here's how you can do that with the OpenStack CLI. You can also do this through the Horizon.

openstack application credential create --description "Kubernetes" kubernetes

Create configuration file​

On an environment with access to Kubernetes CLI, create the following configuration file. Name the file cloud.conf. This file is used by the OpenStack cloud provider to authenticate with your OpenStack.

  • auth-url: You can find this under "API Access" within your project. It's labeled as "Identity".
  • floating-network-id: The ID of the external network floating IPs are issued on.
  • subnet-id: The ID of the subnet within your network.
[Global]
auth-url=http://192.128.1.15:5000
application-credential-id=
application-credential-secret=

[LoadBalancer]
use-octavia=true
floating-network-id=
subnet-id=

Create a secret​

Create a secret with the configuration file.

kubectl create secret -n kube-system generic cloud-config --from-file=cloud.conf

Deploy provider​

Deploy the OpenStack Cloud Controller Manager​

kubectl apply -f https://raw.githubusercontent.com/kubernetes/cloud-provider-openstack/master/manifests/controller-manager/cloud-controller-manager-roles.yaml
kubectl apply -f https://raw.githubusercontent.com/kubernetes/cloud-provider-openstack/master/manifests/controller-manager/cloud-controller-manager-role-bindings.yaml
kubectl apply -f https://raw.githubusercontent.com/kubernetes/cloud-provider-openstack/master/manifests/controller-manager/openstack-cloud-controller-manager-ds.yaml

Wait for pods to be ready​

The provider will attempt to create a pod on each of your control plane nodes. We'll watch for them to be ready.

kubectl get pods -n kube-system -w

Output:

NAME                                       READY   STATUS      RESTARTS      AGE
openstack-cloud-controller-manager-4qrdb   1/1     Running     0             2d1h
openstack-cloud-controller-manager-hjhn8   1/1     Running     0             2d1h
openstack-cloud-controller-manager-jqpbk   1/1     Running     0             2d1h

Troubleshooting​

  • If pods do not automatically deploy on control plane nodes

    During our testing with RKE1, we found that the pods did not deploy on the control plane nodes. We had to manually update the nodeSelector in the DaemonSet to add the label of our control plane nodes.

    The default nodeSelector is: node-role.kubernetes.io/control-plane: ""

    node-role.kubernetes.io/controlplane: "true"

    Note that control-plane becomes controlplane.

    Edit the DaemonSet with the following command to update the nodeSelector:

    kubectl edit daemonset/openstack-cloud-controller-manager -n kube-system

Verify​

We'll verify the functionality of your cloud provider by creating a Kubernetes load balancer service. This will create a load balancer in OpenStack and assign a floating IP to it.

Deploy a simple application​

kubectl run hostname-server --image=lingxiankong/alpine-test --port=8080

Create a load balancer service​

kubectl expose pod hostname-server --type=LoadBalancer --target-port=8080 --port=80 --name hostname-server

Watch for the service to be ready​

When the service is ready, you'll see the EXTERNAL-IP field populated with the floating IP of your load balancer. This may take a few minutes.

kubectl get svc hostname-server -w

Output:

NAME              TYPE           CLUSTER-IP      EXTERNAL-IP      PORT(S)        AGE
hostname-server   LoadBalancer   10.43.228.192   <pending>          80:31835/TCP   2d1h

Troubleshooting:

  • If your service is stuck in a pending state, but your load balancer is created in OpenStack, you can assign a floating IP to the load balancer manually. The service will be automatically updated with the external IP.

Verify the load balancer​

curl http://<floating-ip>

Output:

hostname-server

What's Next​

Now that you've configured, deployed, and verified your cloud provider, you can set up other cloud provider features. Cinder, Barbican, and Octavia are all supported by the cloud provider.

We'll cover how to configure them in our next Kubernetes guides. For now, please see the Cloud Provider OpenStack.