StorageOS clusters represent groups of nodes which run a common distributed control plane, and aggregate their storage into one or more pools.

Typically, a StorageOS cluster maps one-to-one to a Kubernetes (or similar orchestrator) cluster, and we expect our container to run on all worker nodes within the cluster that will consume or present storage.

Clusters use etcd to maintain state and manage distributed consensus between nodes. We offer a choice between an internally managed etcd suitable for test installations or the ability to interface with an external etcd, suitable for production deployments. We recommend the use of external etcd when production or production like workloads will be deployed on StorageOS.