The Vagrantfile and scripts will build a Kubernetes cluster with a single Control Plane
and 2 Nodes. You can increase the nodes in the Vagrantfile if need be. There are many
ways to install kubernetes, kubeadm init is used here
Vagrant script will do the following.
- Build the required Virtual Machines using Virtualbox
- Update the ubuntu linux and prepare for kubernetes
- Install all kubernetes and docker software
- Configure & Build kubernetes Control Plane
- Configure nodes and join cluster
- Install Weave CNI
If you are going to work with Kubernetes, you must familiarise yourself with the kubernetes documentation - which is excellent.
https://kubernetes.io/docs/home/
Read the install scripts in ubuntu/vagrant for more information on the install process
If you are looking to learn kubernetes and possibly do the Certified Kubernetes Admin exam, you might find this helpful. It's stil very much a work in progress.
This files and code in this repository have only been tested on a linux ubuntu & Mac OS host. You need to install the following software on your host. Apologies to Windows users, I don't have access to windows.
git:
https://git-scm.com/downloads
vagrant:
https://www.vagrantup.com/docs/installation
Oracle virtualbox
https://www.virtualbox.org
If you have never used these before, after you have completed your install, check that you can access the commands from the command line
git --version
vboxmanage --version
vagrant version
Apologies to Windows users, I'm not sure how the command line works, dont have windows
os to test
git clone https://github.com/minkley/kubestuff.git
cd kubestuff # change to the kubestuff directory
vagrant validate # Validate the downloaded Vagrantfile
Vagrant file validated successfully.
#NOTE. Bridged adapter selection.
When you first run vagrant up kmaster it is most likely that you will be prompted
to manually select the bridged adapter. Make a note of the adapter name in the
prompt and replace "enp66s0" with the correct adapter for your system, in the following line of the Vagrantfile
# Create a public network, which generally matched to bridged network.
# Bridged networks make the machine appear as another physical device on
# your network.
config.vm.network "public_network",bridge: "enp66s0"
^^^^^^^
The vm names configured in the Vagrantfile are:
- kmaster
- knode01
- knode02
Build the kmaster first before the nodes. The kmaster build writes the output of the kubeadm init command which is needed for the nodes to join the cluster
Build kmaster first
vagrant up kmaster
Build the nodes after you have successfully built kmaster
vagrant up knode01 knode02
Each vm is created with user/password vagrant. PLEASE DO NOT USE IN PRODUCTION.
vagrant ssh kmaster
You can also run ssh vagrant@kmasterIP
Choose which CNI you want to install. The default is Calico
The script - install-CNI-Calico.sh is sourced from Vagrantfile
export KUBECONFIG=/etc/kubernetes/admin.conf
kubectl apply -f https://docs.projectcalico.org/manifests/calico.yaml
The script, install-CNI-Weave.sh is run during the build on kmaster.
export KUBECONFIG=/etc/kubernetes/admin.conf
kubectl apply -f "https://cloud.weave.works/k8s/net?k8s-version=$(kubectl version | base64 | tr -d '\n')"
After you have joined all the knodes, you can run a few commands to check the status of the cluster.
kubectl get nodeskubectl cluster-infokubectl get pods --all-namespaces
List all the nodes.
vagrant@kmaster:~$ kubectl get nodes
NAME STATUS ROLES AGE VERSION
kmaster Ready control-plane,master 7m16s v1.22.1
knode01 Ready <none> 3m55s v1.22.1
knode02 Ready <none> 59s v1.22.1
Show the clusterinfo
vagrant@kmaster:~$ kubectl cluster-info
Kubernetes control plane is running at https://192.168.56.2:6443
CoreDNS is running at https://192.168.56.2:6443/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
List pods in all namespaces
vagrant@kmaster:~$ kubectl get pods -A
NAMESPACE NAME READY STATUS RESTARTS AGE
kube-system coredns-78fcd69978-pdrbx 1/1 Running 0 8m50s
kube-system coredns-78fcd69978-qq9mb 1/1 Running 0 8m50s
kube-system etcd-kmaster 1/1 Running 0 9m4s
kube-system kube-apiserver-kmaster 1/1 Running 0 9m2s
kube-system kube-controller-manager-kmaster 1/1 Running 0 9m4s
kube-system kube-proxy-fpwjh 1/1 Running 0 2m49s
kube-system kube-proxy-p72q2 1/1 Running 0 8m51s
kube-system kube-proxy-txfgk 1/1 Running 0 5m45s
kube-system kube-scheduler-kmaster 1/1 Running 0 9m2s
kube-system weave-net-4pxpg 2/2 Running 1 (8m35s ago) 8m51s
kube-system weave-net-h5l2v 2/2 Running 0 5m45s
kube-system weave-net-mbdlk 2/2 Running 0 2m49s
kubectl run nginx --image=nginx
kubectl run nginx1 --image=nginx
vagrant@kmaster:~$ kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx 1/1 Running 0 30s 10.32.0.2 knode01 <none> <none>
nginx1 1/1 Running 0 25s 10.32.0.2 knode02 <none> <none>
Delete the pods
vagrant@kmaster:~$ kubectl delete pods nginx nginx1
pod "nginx" deleted
pod "nginx1" deleted
Kelsey Hightowers kubernetes the hard way is an excellent link for getting started and is recommended by the Linux Foundation CKA training. Kubernetes the hard way
To get started on installing Kubernetes, take a look here
Kubernetes API Server validates and configures data for pods, services, replication etc.
The API server is accessed by REST. more info here
- Run the following command on your master create command completion for kubectl
kubectl completion bash >/etc/bash_completeion.d/kubectl
- I have found during developing with VM's, creating and deleting VM's can sometimes cause strange behaviour with Virtualbox or your vms. When this happens before you rebuild or delete anythin, try rebooting your host. This usually worked for me.