Create a cluster
This content is for 0.1.0. Switch to the latest version for up-to-date documentation.
Workload clusters are FleetClusterJoin (FCJ) objects with this lifecycle:
Pending → ScriptIssued → Enrolled → Validated → Provisioning → Ready
- Create the FCJ in an authenticated console session or with an authorized API client. Select a registry site, supported topology, node addresses, and any optional storage or edge settings.
- The controller allocates overlay addresses and issues a bootstrap bundle.
- Retrieve that bundle only through the authenticated console or authorized
API flow, then run
sudo ./bootstrap.shon the designated first control node. - After the cluster validates, request provisioning from the console. The hub joins it to OCM, applies the release platform resources, and marks it ready once the managed cluster is available.
Bootstrap bundle handling
Section titled “Bootstrap bundle handling”The bundle contains enrollment material, WireGuard configuration, and cluster-bootstrap credentials. Treat it as sensitive:
- do not publish, email, commit, or retain it in shared storage;
- transfer it only to the intended bootstrap host over an approved channel; and
- revoke or recreate the FCJ if its contents may have been disclosed.
Provisioning does not install a workload provider unless its FCJ install intent
was selected. v0.1.0 FCJ provider kinds are k8s, helm, and kubevirt only.
Continue with Install providers or
read the Cluster join lifecycle.