Cluster Lifecycle¶
The Cluster Lifecycle describes how Digital Rebar Provision manages groups of machines as a coordinated unit. Clusters have their own lifecycle and workflows that operate independently from the individual machines they contain.
Unlike the Machine Lifecycle (which provisions a single machine), the Cluster Lifecycle coordinates multi-machine deployments — provisioning, resizing, and destroying entire groups through cluster-level workflow chains.
How Clusters Differ from Machine Lifecycle¶
| Aspect | Machine Lifecycle | Cluster Lifecycle |
|---|---|---|
| Subject | One machine | A group of machines |
| Workflows | universal-discover, universal-start |
universal-cluster-provision, universal-cluster-destroy |
| Allocation | Direct | Via Resource Broker from a Pool |
| State tracking | Machine object | Cluster object + per-machine cluster/tags profile |
Cluster Lifecycle Phases¶
- Provision — A Resource Broker allocates machines from a Pool; the cluster workflow drives each machine through its individual pipeline while tracking group state
- Operation — The cluster can be placed in WorkOrder Mode to manage ongoing operations (Kubernetes app deployments, cluster resizing, etc.)
- Destroy — The cluster workflow deprovisions machines and returns them to their Pool via the Resource Broker
In This Section¶
- Cluster Architecture — How clusters work, the Cluster object, workflow chains, and relationship to Resource Brokers and Pools
- Cluster Workflows — The provision/destroy pipeline structure and how it reuses machine lifecycle patterns
See Also¶
- Execution Engine: Resource Brokers — How machines are allocated for clusters
- Execution Engine: Pooling — The machine pools that Resource Brokers draw from
- Provisioning Pipeline — The workflow engine driving cluster operations
- Operator Guide: Cluster Operations