All the code for this chapter (and other chapters) is available at https://github.com/param108/kubernetes101 check the directory 014

Changing tack a bit, today I am playing with another Kubernetes simulation environment called kind.

What is it ?

Kind allows you to run kubernetes on “nodes” which are docker images. You can configure the number of nodes and then use kubectl as if on a normal cluster.

Kind configuration is pretty straight forward on ubuntu 18.04 and there is also brew install for Mac folks.

curl -Lo ./kind https://github.com/kubernetes-sigs/kind/releases/download/v0.7.0/kind-$(uname)-amd64
chmod +x ./kind

Creating a cluster is as easy as

$ kind create cluster
Creating cluster "realkube" ...
 ✓ Ensuring node image (kindest/node:v1.17.0) đŸ–ŧ
 ✓ Preparing nodes đŸ“Ļ  
 ✓ Writing configuration 📜 
 ✓ Starting control-plane 🕹ī¸ 
 ✓ Installing CNI 🔌 
 ✓ Installing StorageClass 💾 
 ✓ Joining worker nodes 🚜 
Set kubectl context to "kind-kind"
You can now use your cluster with:

kubectl cluster-info --context kind-kind

Would have been cool to call it kindly no ? For example kindly create cluster 🙂

Multiple Nodes

The really cool thing here is that Kind supports multiple nodes. To create multiple nodes you need to create a config file.

kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
# One control plane node and three "workers".
#
# While these will not add more real compute capacity and
# have limited isolation, this can be useful for testing
# rolling updates etc.
#
# The API-server and other control plane components will be
# on the control-plane node.
#
# You probably don't need this unless you are testing Kubernetes itself.
nodes:
- role: control-plane
- role: worker
- role: worker
- role: worker

$ kind create cluster --name realkube --config config.yml 
Creating cluster "realkube" ...
 ✓ Ensuring node image (kindest/node:v1.17.0) đŸ–ŧ
 ✓ Preparing nodes đŸ“Ļ đŸ“Ļ đŸ“Ļ đŸ“Ļ  
 ✓ Writing configuration 📜 
 ✓ Starting control-plane 🕹ī¸ 
 ✓ Installing CNI 🔌 
 ✓ Installing StorageClass 💾 
 ✓ Joining worker nodes 🚜 
Set kubectl context to "kind-realkube"
You can now use your cluster with:

kubectl cluster-info --context kind-realkube

See the 4 packages next to nodes ?. It takes a couple of minutes for all the nodes to come to ready state.

$ kubectl get nodes
NAME                     STATUS     ROLES    AGE     VERSION
realkube-control-plane   Ready      master   2m14s   v1.17.0
realkube-worker          NotReady   <none>   94s     v1.17.0
realkube-worker2         NotReady   <none>   94s     v1.17.0
realkube-worker3         NotReady   <none>   94s     v1.17.0

And finally, after a while, this is the output of get nodes.

$ kubectl get nodes
NAME                     STATUS   ROLES    AGE     VERSION
realkube-control-plane   Ready    master   2m27s   v1.17.0
realkube-worker          Ready    <none>   107s    v1.17.0
realkube-worker2         Ready    <none>   107s    v1.17.0
realkube-worker3         Ready    <none>   107s    v1.17.0

Node Affinity

Kubernetes allows you to suggest which nodes a pod can run on. One way to do this is to use spec.nodeSelector.

NodeSelector

Lets use our simple ubuntuPod from chapter 5 to test this out on kind. The Pod spec is available in the directory 014/ubuntuPod.yml. It is reproduced here.

apiVersion: "v1"
kind: Pod
metadata:
  name: ubuntu-pod
  labels:
    app: ubuntu-pod
    version: v5
    role: backend
spec:
  containers:
  - name: ubuntu-container
    image: ubuntu
    command: ["/bin/bash"]
    args: ["-c", "while [ \"a\" = \"a\" ]; do echo \"Hi\"; sleep 5; done" ]

There are 4 nodes to select from, lets choose the node realkube-worker.

Label the node

The first step is to label the node. To label a node, you use the following command

kubectl label nodes <node-name> <label-key>=<label-value>

Lets label realkube-worker with quality = best

$ kubectl label nodes realkube-worker quality=best
node/realkube-worker labeled

$ kubectl get nodes/realkube-worker -o yaml
apiVersion: v1
kind: Node
metadata:
  annotations:
    kubeadm.alpha.kubernetes.io/cri-socket: /run/containerd/containerd.sock
    node.alpha.kubernetes.io/ttl: "0"
    volumes.kubernetes.io/controller-managed-attach-detach: "true"
  creationTimestamp: "2020-04-01T12:12:41Z"
  labels:
    beta.kubernetes.io/arch: amd64
    beta.kubernetes.io/os: linux
    kubernetes.io/arch: amd64
    kubernetes.io/hostname: realkube-worker
    kubernetes.io/os: linux
    quality: best
  name: realkube-worker

You can see the labels under metadata

Set NodeSelector in the pod

Add a NodeSelector section to the pod spec.

apiVersion: "v1"
kind: Pod
metadata:
  name: ubuntu-pod
  labels:
    app: ubuntu-pod
    version: v5
    role: backend
spec:
  containers:
  - name: ubuntu-container
    image: ubuntu
    command: ["/bin/bash"]
    args: ["-c", "while [ \"a\" = \"a\" ]; do echo \"Hi\"; sleep 5; done" ]
  nodeSelector:
    quality: best

Lets apply this podspec and check if its node is chosen correctly.

kubectl get pods/ubuntu-pod -o wide gives a nice output with the necessary node details.

$ kubectl apply -f ubuntuPod.yml 
pod/ubuntu-pod created

$ kubectl get pods/ubuntu-pod  -owide
NAME   READY   STATUS    RESTARTS   AGE     IP   NODE    NOMINATED NODE   READINESS GATES
ubuntu-pod   1/1     Running   0          4m38s   10.244.3.2   realkube-worker   <none>           <none>

Its a bit hard to see in the above output, but NODE is realkube-worker.

Multiple nodes can have the same label and kubernetes will choose one of them to schedule the pod.

There is a better way to program affinity to particular nodes, we will look at those in the next post.

Delete a cluster

To delete the cluster and nodes, just do 

kind delete cluster --name=realkube

Learnings

Kind allows you to simulate multi node clusters on your laptop

You can setup Kind by following instructions here.

You can assign a Pod to a particular node, using a spec.nodeSelector.

Conclusion

In this post we have mostly looked at Kind and looked at one way to suggest the node on which a pod should run. In the next post we will look at reasons why we would need to do this and one other way to do this, namely – Affinity and AntiAffinity.