What is this project all about?

In essence this project is used to make it easier to set up a kubernetes cluster on Alpine Linux.

'Why use Alpine?', you ask. Alpine is a linux distro that focuses on 2 things, security and distro size. By having a smaller distro, you tend to have a lot less packages and libraries installed. This means there is a less likelyhood that one of them have an unknown exploitable vulnerability.

Also by making use of a lightweight OS, you have more resources dedicated to running your apps. Less is more.

Furture prospects

I would like to find a suitable method to use the minirootfs version of Alpine instead of the the standard Alpine ISO, because the minirootfs is around 3.5MB and standard Alpine ISO is roughly 200MB. The only thing missing for the minirootfs is a boot functionality.

I want to find a way to build an compact ISO alpine image that is built around using Kubernetes in a secure datacenter environment. This would include seamless and simple kernel updates, quick cluster setup and the capabilities to monitor your physical machines.

Such as your Network packets, latency, CPU usage, RAM, storage and various other common useful statistics. Unfortunately, for now we'll have to settle with humble beginnings, which is a script that needs a little bit of work.

Alpine releases

Alpine Download Page: [ https://www.alpinelinux.org/downloads/ ]

I reccomend the standard x86_64 version, even though it might be an old version the init.sh script will update to the latest stable version.

Install alpine

Localhost login: root
Run setup-alpine

Example of Alpine setup

Select keyboard layout [none]: us
Select variant (or 'aboort'): us
Enter system hostname [localhost]: ENTER
Available interfaces are: eth0 eth1
Which one do you want to initialize? [eth0]: ENTER
Ip address for eth0? [dhcp]: ENTER
Do you want any manual network configuration? (y/n) [n]: ENTER
New Password: root
Retype Password: root
Which Timezone are you in? ('?' for list) [UTC]: Europe/Berlin
HTTP/FTP proxy URL? [none]: ENTER
Which NTP client to run? ('busybox','openntpd','chrony' or 'none') [chrony]: busybox
Press q
Enter mirror number (1-72) or URL to add (or r/f/e/done) [1]: r
Setup a user? (enter a lower-case loginname, or 'no') [no]: ENTER
Which server? ('openssh','dropbear', or none) [openssh]: ENTER
Allow root ssh login? ('?' for help) [phrohibit-password]: ENTER
Enter ssh key or URL for root (or 'none') [none]: ENTER
Available disks are: sda
Which disks would you like to use? [none]: sda
How would you like to use it? ('sys', 'data', 'crypt', 'lvm' or '?') [?]: sys
WARNING: Erase above disk(s) and continue? (y/n) [n]: y

NOTE: Don't forget to unmount iso and reboot.

reboot

Download Kubernetes setup script

Unfortunately to get the init.sh script you have to type out one of 2 options:
Option 1: Download the raw file

wget https://raw.githubusercontent.com/resphantom/alpine-k8/master/init.sh

Option 2: Install git and clone the repository

apk add git
git clone https://github.com/resphantom/alpine-k8.git
cd alpine-k8-cluster

Give the init.sh permission to execute

chmod +x init.sh

Install Kubernetes

Note: This script has to be run for both master and worker nodes You can set the kubernetes version manually by setting the KUBE_VERSION variable

export KUBE_VERSION=1.27

Run the folloing command
Note: init.sh will copy and rename itself to /bin/kubecom

./init.sh init

Register master node

Automated method

kubecon generate-cluster

Manual method

Update hostname

hostname master-1
echo "master-1" > /etc/hostname

Forward IP address

echo 1 > /proc/sys/net/ipv4/ip_forward

NOTE: In the following command you can add the --ignore-preflight-errors=all flag to bypass preflight checks for machines with less than recommended resources, however the Kubernetes cluster might not fully install correctly.

Recommended resources are 2 CPU and 2 GB RAM

Initialise Kubernetes cluster master node.

kubeadm init --pod-network-cidr=10.244.0.0/16 --node-name=$(hostname)

Setup Kubernetes cli config by replacing existing config with a symlink to the admin config

rm /root/.kube/config || mkdir ~/.kube
ln -s /etc/kubernetes/admin.conf /root/.kube/config

Install a CNI controller, we can use flannel for simplicity

kubectl apply -f https://raw.githubusercontent.com/flannel-io/flannel/master/Documentation/kube-flannel.yml

You should now be able to use the custom join command in the other kuernetes machine setup by generating the join command

kubeadm token create --print-join-command

Confirm setup

kubectl get all -n kube-system

Register worker node

On a master node run the following join command

kubeadm token create --print-join-command

On the worker node update the hostname

hostname worker-1
echo "worker-1" > /etc/hostname

NOTE: still looking for a simple way to do this, but for now this is the way
Copy or retype over the uniquely generated command from the master node to the worker node, example of a kubernetes join command below:

kubeadm join 10.0.0.150:6443 --token xunjoc.yx2m65r8inhxph9i --discovery-token-ca-cert-hash sha256:e38dd277fe1143771dfe17261d9862e5313d1cdf3922ea86f8f73b6c0a515798