IPFS Operator

This operator is still heavily in progress.

Running

This operator can be deployed either with or without OLM installed.

With OLM

operator-sdk run bundle quay.io/redhat-et-ipfs/ipfs-operator-bundle:v0.0.1 -n ipfs-operator-system

Without OLM

make deploy

Deploying an IPFS cluster

The value for URL must be changed to match your Kubernetes environment. The public bool defines if a load balancer should be created. This load balancer allows for ipfs gets to be done from systems outside of the Kubernetes environment.

apiVersion: cluster.ipfs.io/v1alpha1
kind: Ipfs
metadata:
  name: ipfs-sample-1
spec:
  ipfsStorage: 2Gi
  clusterStorage: 2Gi

Once the values match your environment run the following.

kubectl create -n default -f ifps.yaml

Running in KIND

An easy way to test and modify changes to the operator is by running it in a local KIND cluster. To bootstrap a KIND cluster, you can run hack/setup-kind-cluster.sh, which will install all of the required components to operate an IPFS cluster.

To deploy the operator in this repository into the cluster, you can run hack/run-in-kind.sh which will build the source code and inject it into the cluster. If you make subsequent changes, you will need to re-run hack/run-in-kind.sh and kill the previous operator manager by running kubectl delete pod -A -n ipfs-operator-system in order to redploy the updated image.

Testing Local Changes

If you're developing the operator and would like to test your changes locally, you can do this by running the kuttl end-to-end tests with make test-e2e after redploying the operator.