This repo contains the ansible scripts to bootstrap a CAPI cluster using kind as the jump point to get it started.
This is based on Kubernetes Cluster API Provider OpenStack under the getting started link.
You also need to ensure that a CAPI vm image is avaliable, the one used in the NeSI RDC is ubuntu-2204-kube-v1.26.7
terraform
These are the terraform files used to provision the base kind compute instance to bootstrap the CAPI cluster
Make a copy of terraform.tfvars.example and rename it to terraform.tfvars and fill in the requried parameters
cp terraform/terraform.tfvars.example terraform/terraform.tfvars
Inside the terraform/terraform.tfvars file is some user configuration required.
tenant_name = "NeSI_RDC_PROJECT_NAME"
key_pair = "NeSI_RDC_KEYPAIR_NAME"
key_file = "NeSI_RDC_KEYFILE"
kind_flavor_id = "6b2e76a8-cce0-4175-8160-76e2525d3d3d" # balanced1.2cpu4ram
kind_image_id = "1a0480d1-55c8-4fd7-8c7a-8c26e52d8cbd" # Ubuntu 22.04
vm_user = "IMAGE_USER" # Ubuntu is `ubuntu` as an exxample
kind_security_groups = ["6443_Allow_ALL", "SSH Allow All", "default"]
NeSI_RDC_KEYPAIR_NAME is your Key Pair name that is setup in NeSI RDC
NeSI_RDC_KEYFILE is the local location for your ssh key
kind_security_groups need to allow SSH from your ansible host so that the playbooks can run. Ensure you have the following ones created or are the same ports:
6443_Allow_ALL- Allow 6443 to allSSH Allow All- Allow 22 to all
ansible
Make a copy of group_vars/servers/servers.yml.example and rename it to servers.yml and fill in the requried parameters
cp group_vars/servers/servers.yml.example group_vars/servers/servers.yml
Inside the group_vars/servers/servers.yml file is some user configuration required.
---
kubernetes_version: v1.28.5
capi_image_name: rocky-89-kube-v1.28.5
capi_provider_version: v0.8.0
cluster_name: CLUSTER_NAME
cluster_namespace: default
cluster_network: NeSI_RDC_PROJECT_NAME
openstack_ssh_key: NeSI_RDC_KEYPAIR_NAME
cluster_control_plane_count: 1
control_plane_flavor: balanced1.2cpu4ram
cluster_worker_count: 2
worker_flavour: balanced1.2cpu4ram
cluster_node_cidr: 10.30.0.0/24
cluster_pod_cidr: 172.168.0.0/16
bin_dir: /usr/local/bin
clouds_yaml_local_location: ~/.config/openstack/clouds.yaml
CLUSTER_NAME the name for your management cluster
NeSI_RDC_PROJECT_NAME the name of your NeSI RDC project space
NeSI_RDC_KEYPAIR_NAME is your Key Pair name that is setup in NeSI RDC
There are the following CAPI images available
Rocky 8.9
rocky-89-kube-v1.22.17
rocky-89-kube-v1.24.17
rocky-89-kube-v1.26.12
rocky-89-kube-v1.25.16
rocky-89-kube-v1.27.6
rocky-89-kube-v1.28.5
Ubuntu 22.04
ubuntu-2204-kube-v1.22.17
ubuntu-2204-kube-v1.23.4
ubuntu-2204-kube-v1.26.7
ubuntu-2204-kube-v1.27.6
ubuntu-2204-kube-v1.28.3
For management clusters we recommend Kuberenetes version 1.27+
If changing the capi_image_name within servers.yml please also ensure the kubernetes_version matches the same.
Example would be using the image rocky-89-kube-v1.27.6 would mean the kubernetes_version would be v1.27.6
clouds.yaml
You will need to ensure you have downloaded the clouds.yaml from the NeSI RDC dashboard and placed it in ~/.config/openstack/clouds.yaml
It is recommended that you use Application Credentials rather then your own credentials.
deployment.sh
To bootstrap the new CAPI cluster run the command ensuring you supply the following variables as envars
NeSI_RDC_KEYFILE_LOCATION location to your local keyfile
VM_USERNAME is the username for the kind image
export TF_VAR_key_file="NeSI_RDC_KEYFILE_LOCATION" export TF_VAR_vm_user="VM_USERNAME"
./deployment.sh bootstrap
This will run 3 playbooks.
ansible-playbook setup-infra.yml -e operation=create
This first creates the kind compute instance inside the NeSI RDC
ansible-playbook -i host.ini ansible-kind.yml -u ubuntu --key-file '~/.ssh/id_flexi'
Runs the ansible play ansible-kind.yml that installs kind onto the compute instance, installs the required dependencies to get it setup as a CAPI cluster, Creates a new workload cluster and promotes that to the management cluster.
You will need to update the values in group_vars/servers/servers.yaml with the main value being clouds_yaml_local_location
ansible-playbook setup-infra.yml -e operation=destroy
The final playbook to run is the terraform destroy, this will tear down the kind instance after the new CAPI Management cluster has been created.
If running this outside of the NeSI RDC then you will need to adjust your values based on your cloud proivder. This also all based on the cloud provider running an openstack base