In this lesson, we will install and use Kubernetes, the Container orchestration from Google.
Make yourself familiar with Kubernetes. Focus on the features and the concepts, described on the following links.
https://kubernetes.io/docs/concepts/
-
What Featuers does Kubernetes provide?
-
Where in our Cloud Stack do you place Kubernetes?
Cloud Stack Example Deployment Tool Application Component Mediawiki ? Containers Docker ? Virtual Resource Instance m1.small ? Cloud Platform OpenStack -
Cleanup your bwCloud tenant. Next, start four virtual machines from image Ubuntu 16.04, put them into your private network and add floating IPs. One of these four VMs will be your aster, the other three will be worker nodes.
We will now follow these guides to install Kubernetes:
- https://kubernetes.io/docs/setup/independent/install-kubeadm/
- https://kubernetes.io/docs/setup/independent/create-cluster-kubeadm/
Install Docker and Kubernetes on all your virtual machines:
# install docker
sudo apt-get install -y \
apt-transport-https \
ca-certificates \
curl \
software-properties-common
curl -fsSL https://download.docker.com/linux/ubuntu/gpg \
| sudo apt-key add -
sudo add-apt-repository \
"deb [arch=amd64] https://download.docker.com/linux/ubuntu \
$(lsb_release -cs) \
stable"
sudo apt-get update
sudo apt-get install -y docker-ce
sudo usermod -aG docker ubuntu
# install kubelet and kubeadm
apt-get update && apt-get install -y apt-transport-https
curl -s https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key add -
cat <<EOF >/etc/apt/sources.list.d/kubernetes.list
deb http://apt.kubernetes.io/ kubernetes-xenial main
EOF
apt-get update
apt-get install -y kubectl kubelet kubeadm kubernetes-cni
One node will become the Kubernetes Master. In this host, run kubeadm init as root to initialize
a new kubernetes cluster. The initialization command will print important information for
the cluster:
Your Kubernetes master has initialized successfully!
To start using your cluster, you need to run (as a regular user):
sudo cp /etc/kubernetes/admin.conf $HOME/
sudo chown $(id -u):$(id -g) $HOME/admin.conf
export KUBECONFIG=$HOME/admin.conf
You should now deploy a pod network to the cluster.
Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
http://kubernetes.io/docs/admin/addons/
You can now join any number of machines by running the following on each node
as root:
kubeadm join --token 0a78c2.82b7a3558d6a38a6 192.168.5.6:6443
Run all the kubectl and kubeadm commands as root on all your hosts.
Run the first three commands, which sets the configuration file to access the kubernetes cluster.
The last command kubeadm join will be required to add hosts to the new cluster.
But before we continue to add hosts, we need to set up an overlay network
for Kubernetes, to allow private IP networks for containers across hosts.
kubectl apply -f \
http://docs.projectcalico.org/v2.3/getting-started/kubernetes/installation/hosted/kubeadm/1.6/calico.yaml
Kubernetes defines pods: a component which is executed by starting one or many containers
from a specified image. With Kubernetes, you can list all running Pods:
kubectl get pods --all-namespaces
You should see some kube-* pods, which are core features of Kubernetes, and you should see some calico-* pods, which enable the overlay networking we just installed. If some of the Pods are still in STATUS "Pending", be patient until they are all in "Running" state.
In the mean while, log in to all other hosts an run the kubectl join ... command form the previous output as root. In the master, validate that
the hosts are registered by running kubectl get nodes. You should get the following output:
root@kubernetes-1:~# kubectl get nodes
NAME STATUS AGE VERSION
kubernetes-1 Ready 15m v1.6.6
kubernetes-2 Ready 28s v1.6.6
kubernetes-3 NotReady 6s v1.6.6
kubernetes-4 NotReady 8s v1.6.6
When all nodes are ready, we can start running containers on the kubernetes cluster. The most simple way to start a new container is by running:
kubectl run $nameOfYourPod --image=$imageToStartFrom --port=8080 ...
# e.g. start a web server with 3 replicas:
kubectl run testserver --replicas=3 --image=httpd:latest --port=80
# list the deployment status
kubectl get deployments
# list running pods
kubectl get pods
# show detailed information
kubectl describe pod $podNameFromList
The kubectl describe command shows the private IP address of the containers, and the ClusterIP of the deployment. Inside the virtual machine, the web server is
accessible with this private IP / Cluster IP, but it will not be accessible form the outside. For accessible application deployments,
Kubernetes provides services. To create a new service, we need to expose the pod:
# create service for web servers
kubectl expose deployment testserver --type=NodePort --target-port=80
# list services
kubectl get services
The output of the service list contains the port number, by which the exposed deployment is accessible on any of the hosts. In the example below the web server is accessible by any of the public node IPs and port 32495. An external load balancer then can distribute incoming requests among existing nodes in kubernetes.
NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes 10.96.0.1 <none> 443/TCP 2h
testserver 10.105.254.205 <nodes> 80:32495/TCP 6s
In the previous exercises we started our own registry, built the Dockerfiles as images and pushed them to the registry. For running the Mediawiki application on kubernetes, these steps are required as well.
With the images available, you can run and expose deployments for MySQL and Apache2 with the Mediawiki application on Kubernetes. Try to deploy Mediawiki (Database and Webserver) on Kubernetes.
Use the kubectl and docker command to troubleshoot issues while deploying.