Hello and welcome to the Kubernetes Workshop practical session! Here you'll learn how to create a cluster and deploy Pods to it. This README will be the main guide for attendees on what to do, but please feel free to ask the demonstrator if you have any questions.
Docker Desktop is a GUI for running Docker on your local machine. It also just happens to have a Kubernetes cluster built into it, which we'll be using today. Docker Desktop supports all operating systems so you shouldn't have any trouble installing it!
By default, Docker Desktop doesn't install a Kubernetes cluster for you, you have to tell it to do that.
Now that Kubernetes is installed, you can check that it's running by opening a terminal and running this command:
kubectl get nodesIf everything is installed correctly, you should get an output like this:
NAME STATUS ROLES AGE VERSION
docker-desktop Ready control-plane 2m v1.32.2
For this task, we'll create a new Deployment using the traefik/whoami container as an example. This should introduce you to the basics of applying manifests to a cluster.
It's always a good idea to read the file you're about to apply before you do! For this task, we'll be applying manifests/pods/deployment.yaml.
As you're reading the manifest, see if you can answer these questions:
How many replicas of this Pod are you deploying?
3 replicas.
Which container image are you about to deploy?
[traefik/whoami](https://hub.docker.com/r/traefik/whoami)
What is the name of the Deployment, and thus the name of the Pods, that you're going to deploy?
k8s-workshop
All manifests can be applied with the kubectl command. Normally, you would pass a YAML file from your local machine, but you can also use a URL to a raw text file instead, like so:
kubectl apply -f https://raw.githubusercontent.com/shefcompsoc/k8s-workshop/refs/heads/main/manifests/pods/deployment.yamlYou should now be able to see that your Kubernetes cluster has a new Deployment running. You can check this by running:
kubectl get deploymentsYou should get an output similar to this:
NAME READY UP-TO-DATE AVAILABLE AGE
k8s-workshop 3/3 3 3 22m
So your Deployment has applied correctly, but that doesn't mean your Pods are running yet. To check this, you can run:
kubectl get podsYou should get something like this:
NAME READY STATUS RESTARTS AGE
k8s-workshop-5fdf6657c7-flbk9 1/1 Running 0 31m
k8s-workshop-5fdf6657c7-gjhhp 1/1 Running 0 31m
k8s-workshop-5fdf6657c7-gzwhc 1/1 Running 0 31m
Your Pod names will be different, but as long as you can see three Pods prefixed with k8s-workshop, then you've done it correctly!
By now you've created your Pods and verified that they're all running properly. However, you still can't access them, for that you'll need a Service.
Today, we'll only be using the NodePort Service mode, because the other two require extra setup that is out-of-scope for today's workshop. You can read about them here if you're interested.
Again, always check what you're applying before you apply it! This time we'll be applying manifests/services/nodeport.yaml. See if you can answer these questions:
Which host port will be exposed?
Port 30001 will be exposed on the host.
Which pod port will be exposed?
Port 80 will be exposed.
As before, this is just a simple command:
kubectl apply -f https://raw.githubusercontent.com/shefcompsoc/k8s-workshop/refs/heads/main/manifests/services/nodeport.yamlAgain, run this command to check:
kubectl get servicesYou should get an output like this:
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
k8s-workshop NodePort 10.98.185.92 <none> 80:30001/TCP 3h13m
kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 29d
Don't worry about the kubernetes Service, that's just part of the Docker Desktop installation and not something we need to mess with.
We've now created a new Deployment and Service. This should be enough for us to now test out the container.
To get started, go to 127.0.0.1:30001, you should get an output something like:
Hostname: k8s-workshop-5fdf6657c7-gjhhp
IP: 127.0.0.1
IP: ::1
IP: 10.1.0.157
IP: fe80::b878:2cff:fecb:5e86
RemoteAddr: 192.168.65.3:58718
GET / HTTP/1.1
Host: 127.0.0.1:30001
User-Agent: curl/8.7.1
Accept: */*
Take note of the top line, starting with Hostname:. This line indicates which Pod you're currently accessing, and you should be able to see this change if you reload a few times (you may need to clear browser cache). This is the Service automatically load balancing between our replicas, pretty cool right?
If you're a demonstrator for the practical part of the workshop, please see this guide for some information about common problems and recommended solutions.
Resources used in the making of this workshop: