Upgrade Cilium OSS to Cilium Enterprise (OpenShift 4)

Starting situation

You have an OpenShift 4 cluster with Cilium OSS installed.

While we didn’t notice any downtime during the upgrade, we recommend to expect a downtime of a few seconds.

Prerequisites

kubectl
jq
Optionally yq yq YAML processor (version 4 or higher)
A local setup to compile a cluster catalog, see Running Commodore for details.
olm.source parameter set either in the global defaults or in the tenant repository.

This guide assumes that you’re familiar with making changes to a Project Syn cluster and compiling the cluster catalog to deploy those changes.

Upgrade to Cilium Enterprise

Switch the release to enterprise in your cluster config.

c-cluster-id.yml diff

   cilium:
+    release: enterprise

You can also use yq:

yq eval -i '.parameters.cilium.release = "enterprise"' c-cluster-id.yml

Compile the cluster catalog.
Ensure the ArgoCD app is synced within the cluster.

Either use the GUI or the CLI by looking at the .status.history field:
```
kubectl -nsyn get applications cilium -oyaml | yq '.status.history'
```

Fix ownership of unscoped Helm/OLM resources.

kubectl annotate \
  clusterrole/cilium clusterrolebinding/cilium \
  clusterrole/cilium-operator clusterrolebinding/cilium-operator \
  meta.helm.sh/release-name=cilium-enterprise \
  --overwrite

Cleanup objects the RedHat OLM Operator leaves behind.

kubectl -n cilium get clusterserviceversion -ojson \
  | jq -r '.items[] | select(.spec.displayName == "Cilium") | .metadata.name' \
  | xargs kubectl -n cilium delete clusterserviceversion

Ensure the Cilium OSS config file can be deleted.

kubectl -n cilium patch ciliumconfig cilium -p '{"metadata":{"finalizers":null}}' --type=merge