The Fn project is an open source, container native, and cloud agnostic serverless platform. It’s easy to use, supports every programming language, and is extensible and performant.
This chart deploys a fully functioning instance of the Fn platform on a Kubernetes cluster using the Helm package manager.
persistent volume provisioning support in the underlying infrastructure (for persistent data, see below )
Install Helm
Initialize Helm by installing Tiller, the server portion of Helm, to your Kubernetes cluster
helm init --upgrade
In order to get a working deployment please pay attention to what you have in your chart values. Here is the bare minimum chart configuration to deploy a working Fn cluster.
If you are installing Fn behind an ingress controller, you'll need to have a single DNS sub-domain that will act as your ingress controllers IP resolution.
Important: An ingress controller works as a proxy, so you can use the ingress IP address as an HTTP proxy:
curl -x http://<ingress-controller-endpoint>:80 api.fn.internal
{"goto":"https://github.com/fnproject/fn","hello":"world!"}
In order to natively expose the Fn services, you'll need to modify the Fn API, Runner, and UI service definitions:
fn_api
node values, modify fn_api.service.type
from ClusterIP
to LoadBalancer
fn_lb_runner
node values, modify fn_lb_runner.service.type
from ClusterIP
to LoadBalancer
ui
node values, modify ui.service.type
from ClusterIP
to LoadBalancer
In an Fn deployment with LoadBalancer service types, you'll need 3 DNS names:
api.fn.mydomain.com
)lb.fn.mydomain.com
)ui.fn.mydomain.com
)Upon successful deployment, you'll have three public IP addresses -- one for each service. However, the IP address for the API and LB services will be identical since they are exposed as a single service. You'll have two IP addresses, but three DNS names.
Please keep in mind the best way for exposing services is an ingress controller.
Clone the fn-helm repo:
git clone https://github.com/fnproject/fn-helm.git && cd fn-helm
Install chart dependencies from requirements.yaml:
helm dep build fn
The default chart will install fn as a private service inside your cluster with ephemeral storage, to configure a public endpoint and persistent storage you should look at values.yaml and modify the default settings.
To install the chart with the release name my-release
:
helm install --name my-release fn
Note: if you do not pass the --name flag, a release name will be auto-generated. You can view releases by running helm list (or helm ls, for short).
Please ensure that your ingress controller is running and has a public-facing IP address.
An ingress controller acts as a proxy between your internal and public networks.
Therefore in order to talk to your Fn Deployment, you'll need to set the HTTP_PROXY
environment variable or use cURL like so:
curl -x http://<ingress-controller-endpoint>:80 api.fn.internal
{"goto":"https://github.com/fnproject/fn","hello":"world!"}
Assuming your release is named my-release
:
helm delete --purge my-release
The command removes all the Kubernetes components associated with the chart and deletes the release.
For detailed configuration, please see default chart values.
Fn persists application data in MySQL. This is configured using the MySQL Helm Chart.
By default this uses container storage. To configure a persistent volume, set mysql.*
values in the chart values to that which corresponds to your storage requirements.
e.g. to use an existing persistent volume claim for MySQL storage:
helm install --name testfn --set mysql.persistence.enabled=true,mysql.persistence.existingClaim=tc-fn-mysql fn