There are many approaches to deploying helm charts into a kubernetes cluster such as
- Plain helm install
- Gitops (eg fluxcd or argocd)
- Ansible kubernetes modules
Any of the above approaches are perfectly valid. For the sake of simplicty, this article will describe an approach using the Ansible kubernetes modules
There are a number of required dependencies which must be installed before proceeding.
Required:
- ansible
- Helm client
- Pip3
- venv
- jq
- jmespath
- A running kubernetes cluster
- Jetstack certmanager installed and configured correctly on the kubernetes cluster
- DNS for the hostnames configured which will be used for the ingress objects
- A previously downloaded kubeconfig file and the path to the file
- A domain name
- A google recaptche linked to the domain and the google recaptcha private key
- A vatlayer key and vatlayer api key
Optional:
- An installed linkerd2 service mesh
- An openexchanges account and the api key
- A braintree account and braintree private key
The script below will install the required dependencies and activate a virtual environment from within which the ansible playbooks can be run
sudo apt-get install python3-pip
source example/install_deps.sh
install_k8s_venv
install_helm_client
install_kubectl
activate_venvThe vars/saleor_vars.json file contains the variables which are expected for the deployment by
the ansible playbook. These are primarily secrets. Do not checkin this file into source control
with the real variables.
The install-saleor.yaml can be modified according to your requirements.
The settings for the various chart values.yaml files can be amended depending
on your requirements.
source ../scripts/ci_functions.sh
activate_venv
ansible-playbook \
-e @"vars/certmanager_vars.json" \
./playbooks/certmanager-certs.yaml &&
ansible-playbook \
-e @"vars/saleor_vars.json" \
./playbooks/install-saleor.yaml