demo.md

Run Kubernetes v1.0

Start the VM:

qemu-system-x86_64 \
  -drive file=my-vm-disk.qcow2,format=qcow2 \
  -drive file=user-data.img,format=raw \
  -m 8192 \
  -smp cores=2 \
  -nographic \
  -netdev user,id=usernet,hostfwd=tcp::2222-:22 -device virtio-net,netdev=usernet

Connect to the instance

ssh -p 2222 -o StrictHostKeyChecking=no ubuntu@localhost

lsb_release will indicate Ubuntu 15.04:

ubuntu@ubuntu:~$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu 15.04
Release:        15.04
Codename:       vivid

Check for cgroups v1, it should say tmpfs

stat -fc %T /sys/fs/cgroup/

Run etcd in the background as root and follow the logs, press Ctrl-C when you're ready, this will continue to run in background -

sudo bash -c 'etcd --listen-client-urls http://0.0.0.0:2379 --advertise-client-urls http://localhost:2379 &> /var/log/etcd.log &'; tail -f /var/log/etcd.log

Run kube-apiserver in the background as root and follow the logs, press Ctrl-C when you're ready, this will continue to run in background -

sudo bash -c 'kube-apiserver --etcd-servers=http://localhost:2379 --service-cluster-ip-range=10.0.0.0/16 --bind-address=0.0.0.0 --insecure-bind-address=0.0.0.0 &> /var/log/kube-apiserver.log &'; tail -f /var/log/kube-apiserver.log

With kube-apiserver running, we will be able to see this via cluster-info -

kubectl cluster-info

Show kubectl configuration -

kubectl config view

If we check with kubectl get nodes, although it will connect to the API server, currently we will have no nodes

kubectl get nodes

Run the kubelet in the background as root and follow the logs, you're looking for a message similar to "Successfully registered node ubuntu", press Ctrl-C when you're ready, this will continue to run in background. This will register this node with the api-server. Click "Reject" when you get a popup with the title "Authorize Cloud Shell" this is not required, kubelet is checking if the node is running on a cloud provider vm.

sudo bash -c 'kubelet --api-servers=http://localhost:8080 --pod-infra-container-image="registry.k8s.io/pause:0.8.0" &> /var/log/kubelet.log &'; tail -f /var/log/kubelet.log

If we show nodes, we will now see one node -

kubectl get nodes

Run the kube-scheduler in the background as root and follow the logs (expect to see no output),press Ctrl-C when you're ready, this will continue to run in background -

sudo bash -c 'kube-scheduler --master=http://localhost:8080 &> /var/log/kube-scheduler.log &'; tail -f /var/log/kube-scheduler.log

Run the kube-controller-manager in the background as root and follow the logs,press Ctrl-C when you're ready, this will continue to run in background -

sudo bash -c 'kube-controller-manager --master=http://localhost:8080 &> /var/log/kube-controller-manager.log &'; tail -f /var/log/kube-controller-manager.log

Run the kube-proxy in the background as root and follow the logs (expect to see no output),press Ctrl-C when you're ready, this will continue to run in background -

sudo bash -c 'kube-proxy --master=http://localhost:8080 &> /var/log/kube-proxy.log &'; tail -f /var/log/kube-proxy.log

Docker Hub will not work, owing to changes in the registry standards, therefore we will manually need to load images. We're going to load nginx:1.7 which at the time is 10 years old, we'll download this and pipe it direct to docker load -

curl -L https://github.com/cloud-native-istanbul/kubernetes-v1.0-lab/raw/main/images/nginx-1.7.tar | sudo docker load

And we'll go back in time and make nginx:1.7 nginx:latest through a manual tag, allowing us to use nginx with no tag in kubernetes -

sudo docker tag nginx:1.7 nginx:latest

We will also require a pause container image which at the time would have been 0.8.0 -

curl -L https://github.com/cloud-native-istanbul/kubernetes-v1.0-lab/raw/main/images/gcr_io_google_containers_pause_0_8_0.tar | sudo docker load

Show the available Docker images -

sudo docker images

Let's try out Kubernetes v1.0. The syntax is different, we'll run 5 pods which in turn, will create a replicationcontroller for us -

kubectl run nginx --image=nginx --replicas=5

Check the pods until they are running -

kubectl get pods

As Docker is the container runtime, you will be able to see the pods running in docker as well -

sudo docker ps -a

We can also expose these as a service -

kubectl expose rc nginx --port=80 --target-port=80

Query the available services -

kubectl get svc

Capture the svc IP -

SVC_IP=$(kubectl get svc nginx -o template --template={{.spec.clusterIP}})

And now you can curl that IP -

curl $SVC_IP

Congratulations, you've successfully used Kubernetes v1.0. Here's to looking forward to another 10 years of Kubernetes!

What actually do we have in this version? Checkout API Timeline