What is the Rancher Cluster Management UI doing differently than in the Helm chart for a Rancher-provisioned cluster? When I provision a RKE2 cluster using this Helm chart (https://github.com/rancher/cluster-template-examples), this does not provision the Harvester CSI driver to the downstream cluster. However, if I provision this cluster the exact same options via the UI, it auto-provisions the Harvester CSI driver within the downstream cluster (and functions perfectly). I’ve looked at the Cluster resources (

kind: Cluster

and

kind: HarvesterConfig

) for both clusters as deployed and can’t spot a delta. The only resource I’ve been able to determine that’s different is a Secret that gets created in the

fleet-default

namespace instantly after you click “Create” in the Rancher Cluster Management UI:

harvesterconfigXXXXX

(last 5 are dynamic) This secret contains a service account and token information of the Harvester cluster needed for the CSI driver to work from the downstream cluster. (it also provisions the service account on the Harvester cluster) How can I replicate the Harvester CSI behavior from the UI via the cluster-template Helm charts? Does anyone know what creates that harvesterconfig secret and what additional resource I need to define in the Helm templates to initiate those resource creations? I’m assuming the UI process is kicking off this script (https://raw.githubusercontent.com/harvester/harvester-csi-driver/master/deploy/generate_addon_csi.sh) but how and what can I do to make that happen via the Helm workflow? If there’s a better room, let me know and I’ll repost in there. Thanks in advance!