Configuration¶

All cluster settings are managed through a single file: cluster.yaml. This page explains every configurable field.

File Structure¶

cluster.yaml has two top-level sections:

cluster:
  # Global cluster settings
  ...

nodes:
  # List of node definitions
  ...

Cluster Section¶

Field	Type	Default	Description
`pod_cidr`	string	`10.244.0.0/16`	CIDR range for pod IP addresses
`service_cidr`	string	`10.96.0.0/16`	CIDR range for Kubernetes service IPs
`calico_version`	string	`v3.31.5`	Calico CNI version to deploy
`box`	string	`ubuntu/jammy64`	Vagrant box image for all VMs

Example¶

cluster:
  pod_cidr: "10.244.0.0/16"
  service_cidr: "10.96.0.0/16"
  calico_version: "v3.31.5"
  box: "ubuntu/jammy64"

Pod CIDR

Changing pod_cidr from the default requires that the value matches the Calico custom resource configuration. The provisioning script handles this automatically.

Nodes Section¶

Each node is defined as a list entry with the following fields:

Field	Type	Required	Description
`name`	string	✅	VM hostname (must be unique)
`ip`	string	✅	Private network IP on `192.168.56.x` subnet
`memory`	integer	✅	RAM allocation in MB
`cpus`	integer	✅	Number of virtual CPUs
`role`	string	✅	Either `"controlplane"` or `"worker"`

Constraints¶

Exactly one node must have role: "controlplane". The Vagrantfile validates this at startup and aborts with an error if violated.
All other nodes must have role: "worker".
Each node must have a unique name and ip.

Control Plane Node¶

- name: "controlplane"
  ip: "192.168.56.10"
  memory: 2048          # Minimum 2048 MB recommended
  cpus: 2               # Minimum 2 CPUs required by kubeadm
  role: "controlplane"

Control Plane Resources

kubeadm requires at least 2 CPUs for the control plane node. Setting cpus: 1 will cause kubeadm init to fail.

Worker Node¶

- name: "node01"
  ip: "192.168.56.11"
  memory: 1536          # Minimum 1024 MB for basic workloads
  cpus: 1
  role: "worker"

Adding Worker Nodes¶

To add a new worker node:

Copy an existing worker block
Change the name (must be unique)
Increment the IP address
Adjust memory and cpus as needed
Run vagrant up

nodes:
  # ... existing nodes ...

  - name: "node04"
    ip: "192.168.56.14"
    memory: 2048
    cpus: 2
    role: "worker"

Removing Worker Nodes¶

To remove a worker node:

Delete or comment out the node entry in cluster.yaml
Destroy the specific VM: vagrant destroy node03 -f

Note

Removing a node from cluster.yaml does not automatically destroy its VM. You must run vagrant destroy <name> separately.

Full Example Configuration¶

cluster:
  pod_cidr: "10.244.0.0/16"
  service_cidr: "10.96.0.0/16"
  calico_version: "v3.31.5"
  box: "ubuntu/jammy64"

nodes:
  - name: "controlplane"
    ip: "192.168.56.10"
    memory: 2048
    cpus: 2
    role: "controlplane"

  - name: "node01"
    ip: "192.168.56.11"
    memory: 1536
    cpus: 1
    role: "worker"

  - name: "node02"
    ip: "192.168.56.12"
    memory: 1536
    cpus: 1
    role: "worker"

  - name: "node03"
    ip: "192.168.56.13"
    memory: 1536
    cpus: 1
    role: "worker"