In this, we'll be setting up a multi node K8s cluster on a local environment (MacOS) using the VMs running on Oracle VirtualBox. It enables us to test our apps locally without having to deploy on remote cloud platforms such as AWS or GCP. The Part 1 will be focusing on setting up the VMs and configuring the network. The cluster will consist of a master node and a worker node.

Required software

Once VirtualBox is installed successfully, follow the steps below to configure Ubuntu based VMs locally.

Step 1: Creating and configuring the master node. Use the following spec when setting up the master node.

4GB of RAM
20GB of disk space
2 CPUs
OS: Linux Ubuntu 64bit

Creating the master node (called masternode)

Allocating 4GB of RAM

Creating and Allocating the Virtual Disk (20GB)

Click on Create to create the master VM (this might take a few seconds).

Step 2: Allocating 2 CPUs.

Once Step 1 is completed we should have the master VM appearing as below. To change number of CPUs, right click on the VM and go to Settings > System > Processor. Then add 2 CPUs.

Step 3: Assigning the bootable Ubuntu 20.04.4 image and starting the VM.

Right click on the VM and go to Settings > Storage. Then Select "Controller: IDE" option and select "Adds optical drive" option as shown below.

Then select Add and select the ISO image from the downloaded location and click on Choose.

Step 4: Launching the VM and installing Ubuntu OS.

The next step is to launch the VM and install Ubuntu on it. First click on start and then make sure the Ubuntu ISO image is selected that we assigned in Step 3.

Then continue with the Ubuntu installation process by selecting Install Ubuntu. The minimal installation would be sufficient in this case where it'll exclude all the unnecessary apps such as Open Office & etc.

Once installed, restart the system.

System will restart successfully.

Step 5: Network Settings.

Make sure to power off the VM and then go to VM settings and then select Network. In Adapter 1, change "Attached to" to Bridge Adapter as shown below. By default this is set to NAT. Leave rest of the settings to default.

Click OK and Power on the VM.

Step 6: Enabling SSH

By default, the SSH may be inactive in Ubuntu OS. SSH is important as it enables us to access these VMs from the local machine in order to continue with the rest of the configuration. Therefore follow the steps below to enable SSH on the VM.

Launch the terminal app in Ubuntu master VM and run the following command.

service ssh status

If SSH is not active, it'll output a message like below.

Unit ssh.service could not be found.

To install SSH, run the following commands.

sudo apt-get update
sudo apt install openssh-server

Check the SSH service again as we did above and we should be able to see it in active state as below.

Step 7: Accessing the VMs via SSH from local.

In order to access the VMs via SSH, we should obtain their IPs first using the command below in the VM.

ifconfig

We can now use the IP address (eg. 192.168.8.134) of the enp0s3 interface and the username provided during the installation to access the VM from local as shown below.

ssh yrol@192.168.8.134

Make sure to shutdown master VM before moving into the next step (i.e. using the "shutdown now" command).

Step 8: Creating the worker node

The next step is to create the worker node. This can be achieved easily by cloning the existing master node as shown below.

Right click on the master VM select the Clone option. Provide a name for the VM and make sure to select Generate new MAC addresses for the MAC Address Policy as shown below.

Click continue and in next step, select the Linked clone option as it consumes minimum amount of space.

Click on Clone and once done we should have our new VM created.

Power up the new VM and make sure it can be accessed via SSH as we did in Step 6 for the master VM by grabbing its IP address via ifconfig command (username and password will be the same as master since this master VM clone).

Step 9: Setting up unique hostnames

It is important set up unique hostnames for each VM. Currently we have the same hostname in both VMs. To do this, SSH into both VMs as root user.

The following needs to be repeated in both master and worker nodes using a different hostname (the example below is for the master VM).

Edit the hostname file as below using the vi or vim editor.

vi /etc/hostname

Change the existing hostname to something unique. In this case, it's been changed as masternode. Save the changes (i.e. wq!).

Next change the hosts file by providing the same hostname as above.

vi /etc/hosts

Save the changes (i.e. wq!) and shutdown the VM. Repeat the same for the other VMs.

Step 10: Assigning unique / static IPs to the VMs.

Currently the IP addresses of these VMs are assigned through the home router network (via the bridged adapter). Hence, they dynamic and could change at any time. The solution is to create a host only network and then assign static IPs to all VMs manually. These IPs will be used for internal communication within the cluster between nodes. In Part 2, these static IPs will be referenced during the network configuration step.

Select Tools > Network in VirtualBox (global settings).

Make sure it already has a Virtual Host option available as shown below with an IP range assigned to it. If not go ahead and create one. Also make sure the DHCP Server option is disabled and Configure Adapter Manually option is selected.

Next, we need to utilise the above host on each VM of the cluster. To do that, go to the Settings option of VM and select Network. Then Select Adapter 2 (Adapter 1 has already been utilised for the VM's internet connectivity), and select the following options.

Enable Network Adapter (checked)
Attached to: Host-only adapter (selected from the list)
Name: vboxnet0 (selected from the list)

Leave the rest of the setting to default and click OK.

Next, power up the VMs and SSH. Then run the ifconfig command and you'll see a new interface called enp0s8 without an IP address assigned to it.

Next we need to add a permanent IP for this interface of each VM by following the instructions below.

First we need to modify the file /etc/network/interfaces in order to assign an IP to a particular VM. However, since we're using Ubuntu 20.04 LTS, this does not exist out of the box and we need to install the following package.

sudo apt update
sudo apt install ifupdown net-tools

Once the installation is finished, reboot the system and SSH. Then modify the /etc/network/interfaces file as below by assigning a unique IP to the enp0s8 interface that is within the Virtual Host (vboxnet0) IP range which we configured above (ex: 192.168.56.1/24).

vi /etc/network/interfaces

Save and reboot the system. Check the enp0s8 interface again by running the ifconfig command and we should be able to see the new IP that we just assigned to it as below.

This needs to repeat on the other VMs as well within the cluster by making sure to use unique IPs for enp0s8 interface.

Step 11: Disabling swap

It is a requirement to turn off swap in order for the Kubernetes function properly in all VMs. It can be disabled permanently as below.

sed -i '/ swap / s/^/#/' /etc/fstab

Verify it as below.

cat /etc/fstab

As shown, it is now disabled.

That's it and all the VMs have been set up successfully. In Part 2, we'll be installing and configuring Kubernetes on VMs.