Manually installing Kubernetes 1.10 on Proxmox.

Simple Kubernetes 1.10 install guide.
Updated 6 months ago Published 4 years ago
This guide assumes that you already have an operational Proxmox instance.
Step One - Creating the vms
Now you have your host networking setup your ready to create your virtual machines, for this setup we will be creating a cluster of 3.
The layout will be,
We will be using ubuntu 16.04 for each node, to start ssh into your main Proxmox box and download the Ubuntu ISO the Proxmox template directory, This is so Proxmox can see the ISO to mount it.
cd /var/lib/vz/template/iso
mv ubuntu-16.04.1-server-amd64.iso?_ga=1.150784255.852713614.1480375703 ubuntu-16.04.1-server-amd64.iso
You will now be able to select this ISO when you create your VMS.
Now login to your Proxmox web-ui https://{MAIN_IP}:8006/ with your system credentials
  • Input the hostname, e.g node-01
  • For OS select linux 4.X/3.X/2.6
  • For CD/DVD select from the ISO select menu your downloaded ISO
  • For Hard Disk i recommend 200GB per VM, Space permitting.
  • For CPU use 1 core, Cpu spec permitting.
  • For Memory 4GB, Again memory permitting.
  • For Networking select NAT mode
  • Then confirm.
Now you will also need to add one more network adapter, This adapter will utilize the bridge we created in the previous section.
  • Select the new VM from the "Server View"
  • Find the Hardware option
  • Select "Add" and select "Network Device"
  • We need a "Bridged mode" interface, select bridge vmbr0.
  • Change the "Model" to Intel E1000. These are issues with the standard virtualised network drivers.
  • Now you can turn on your VM.
Install ubuntu as you normally would, By be sure to use the NAT adapter when you install. We will configure the Bridge adapter later.
You will need to repeat this step three times for each VM. Or you can create a template from the first VM you created and clone it three times.
Step Two - Configure VM Network
Once ubuntu is installed you will need to setup the networking for each VM. Connect to node-01 with VNC from the web-ui and login.
Next you will need to configure the adapters, Open up /etc/network/interfaces
vim /etc/network/interfaces
auto ens18
iface ens18 inet dhcp
Your NAT adapter should have already been configured for you, We will now add the Bridged adapter. Add the below to the config.
vim /etc/network/interfaces
auto ens19
iface ens19 inet static
Now restart the networking service
sudo serivce restart networking
Try to ping the Hertzer host
If you are able to ping the host then !! It worked, Your Virtual machine is connected to the main host via the network bridge with its own adapter.
You can confirm this by connecting to the new VM with an SSH tunnel through the hertzer host
ssh -A -t root@{MAIN_IP} ssh -A -t {VM_USER}@{ADDRESSS}
You will need to first input the hertzer hosts password then your new VMS credential. if everything was setup properly then you should be able to SSH into your new VM.
This model uses the NAT adapter to connect to the internet, But you could remove the NAT adapter and just use the private network and use as the network gateway.
You will need to repeat this for each of the VMs assigning each VM a diffrent IP.
You now have a 3 node VM cluster connected via a private network that you can ssh into.
Step Three - Install Kubernetes
Now you will need to install Kubernetes on each of your nodes, So for each VM repeat this process.
Add the Kubernetes repo to your sources list
vim /etc/apt/sources.list.d/kubernetes.list
deb kubernetes-xenial main
Add public key of "Google Cloud Packages Automatic Signing Key"
apt-key adv --keyserver --recv-keys 3746C208A7317B0F
Load in the new repo list
apt-get update
Install base packages
apt-get install -y kubelet kubeadm kubectl kubernetes-cni
Once these packages are installed we can create our base cluster.
We will use node-01 as the master
ssh -A -t root@{MAIN_IP} ssh -A -t {VM_USER}@
On the master we can used the installed kubeadm tool
kubeadm init
This will take several minutes to configure, Once the process finished you will be given a list of detailed that you MUST SAVE.
{VM_USER}@node-01:~# kubeadm init
<master/tokens> generated token: "7fa96f.ddb39492a1874689"
<master/pki> created keys and certificates in "/etc/kubernetes/pki"
<util/kubeconfig> created "/etc/kubernetes/admin.conf"
<util/kubeconfig> created "/etc/kubernetes/kubelet.conf"
<master/apiclient> created API client configuration
<master/apiclient> created API client, waiting for the control plane to become ready
<master/apiclient> all control plane components are healthy after 23.098433 seconds
<master/apiclient> waiting for at least one node to register and become ready
<master/apiclient> first node is ready after 10.034029 seconds
<master/discovery> created essential addon: kube-discovery, waiting for it to become ready
<master/discovery> kube-discovery is ready after 30.44947 seconds
<master/addons> created essential addon: kube-proxy
<master/addons> created essential addon: kube-dns
Kubernetes master initialised successfully!
You can now join any number of machines by running the following on each node:
kubeadm join --token 7fa96f.ddb39492a1874689
The most important command is the kubeadm join command, You need to keep this secret if someone get the token and your master IP they will be able to automatically add nodes to your cluster.
Now install the POD network
kubectl apply -f
Becuase we have a small number of nodes we also want to use our master server as a minion
kubectl taint nodes --all dedicated-
Now you are ready to connect your minions nodes to the master, Assuming you have installed the base packes on node-02 and node-3 simply run,
kubeadm join --token 7fa96f.ddb39492a1874689
This will configure each node.
To check that the nodes are all checked in run,
kubectl get nodes
node-01   Ready     7h
node-02   Ready     7h
node-03   Ready     5h
You should be an output like the one above, Congrats you have a kuberntes cluster.
Step Four - Setup kubectl on your local machine
kubectl config set-credentials ubuntu --username={KUBE_USER} --password={KUBE_PASSWORD}
kubectl config set-cluster personal --server=http://{MAIN_IP}:8080
kubectl config set-context personal-context --cluster=personal --user=ubuntu
kubectl config use-context personal-context
kubectl config set contexts.personal-context.namespace the-right-prefix
kubectl config view