Add RPI Compute Module node to Turing Pi in 10 min

During some of the manipulation of the partition table of my SD card, I ended up screwing up both my SD card and my backup Win32DiskImage backup. Moreover if your SD card is 32G, it takes around 30 minute to restore from backup. Hence the idea to come up with a way to build more resiliency in the cluster. Recreating a node from scratch should not take more than 10 mn. The propose procedure is still rather long because I did not push enough yet what the HypriotOS team, aka build a default SD image where cloud-init does 100% of the initialization work.

Base OS

Flash HypriotOS to SD and reboot Pi.


Flash the SD Card with HypriotOS
Connect Pi through to the cluster switch

Freeze the new node IP

Access the Master PI, run arp -a and find the new IP. Freeze the new IP in the dhcpcd.conf

ssh <masterip>
arp -a
vi /etc/dhcp/dhcpd.conf

Ssh to the new node

Access the node from the Master PI:

docker ps

Basic hostname

Freeze your configuration

sudo apt-get remove --purge cloud-init
sudo apt-get autoremove

Set the host name

sudo vi /etc/hosts
sudo hostnamectl set-hostname <xxx>

Install kubeadm

On Master PI

Firt access the kubemaster node and regenerate a token (if you did not use ttl=0 when using kubeadm init on the master):

kubeadm token create

On New Slave PI

sudo -i

curl -s | apt-key add -
echo "deb kubernetes-xenial main" > /etc/apt/sources.list.d/kubernetes.list
apt-get update && apt-get install -y kubeadm kubectl kubelet

To help during the initialization phase, get kubeadm to download the images onto docker

kubeadm config images pull

Get the node to join the cluster

kubeadm join --token yyyyyy.xxxx --discovery-token-ca-cert-hash sha256:zzzz


  • Will have to come back later and use cloud-init, create a clean & small SD image for Win32DiskImage
  • Will have to create more advanced partition on the SD card.