Deployment scenarios

A Virtuozzo Infrastructure cluster is built from nodes, and each node runs one or more services, such as compute, Virtuozzo Storage, S3, Backup Gateway, iSCSI, or NFS. The first three nodes additionally carry the management role. A node's hardware requirements are the sum of the services it hosts plus headroom; the same server can fill different roles depending on what runs on it.

Select the deployment scenario that matches the planned workload and services. Each scenario page describes service placement, per-node hardware requirements for evaluation and production deployments, storage policies, and a sizing example.

• HCI compute: compute and storage on every node. This deployment scenario introduces the storage policies used throughout the other deployment scenarios.
• HCI compute with S3: adds object storage on a separate tier.
• S3-only: large object-storage deployments with support for multiple storage tiers.
• Compute with external storage: compute that consumes block storage from an external array over iSCSI.

How minimum and recommended are derived

The minimum figure for a node is the sum of the documented service reservations for everything the node runs, plus base OS headroom and a small evaluation-workload allowance, rounded up. The recommended figure targets production load. Storage figures assume a representative node and scale with disk count. If erasure coding is used in the storage cluster, include the per-node storage-client reservation. To size a node that does not match a scenario, sum the reservations directly from CPU and RAM requirements.

The examples use five or six nodes, which is a typical starting point for a small installation. The number of nodes a cluster needs depends on the redundancy mode and failure domain; see the storage policy in each scenario.