Skip to content

Deployment Preparation

Proper deployment planning is essential for ensuring the performance, scalability, and resilience of a simplyblock storage cluster.

Deployment Model

Simplyblock is deployed on Kubernetes and OpenShift and is managed through the Simplyblock Operator and its custom resources. Both disaggregated deployments, with dedicated workers for storage nodes (by default in the same cluster), and hyper-converged deployments, co-located with compute workloads, are supported, as well as a hybrid of both. A wide range of Kubernetes distributions and operating systems is supported. If the same cluster is used to serve and consume the storage, there is no essential difference between the models from a deployment perspective: the choice becomes a question of node selection only. The topologies are described in Deployment Topologies.

Deployment Architectures and Disaster Recovery

Before sizing individual nodes, the overall architecture has to be chosen: how many Kubernetes clusters and storage clusters are involved, where the control plane runs, and whether a hub cluster coordinates disaster recovery between sites.

  • Deployment Architectures: Supported architectures with and without disaster recovery, from a single cluster to a hub with several site clusters, and a decision table.
  • Disaster Recovery Requirements: Hub cluster sizing, Kubernetes versions, site cluster prerequisites, S3, networking, and site labeling for simplyblock Disaster Recovery.

General Information on Requirements

Before installation, key factors such as node sizing, storage capacity, and fault tolerance mechanisms should be carefully evaluated to match workload requirements. This section provides guidance on sizing management nodes and storage nodes, helping administrators allocate adequate CPU, memory, and disk resources for optimal cluster performance.

Additionally, it explores selectable erasure coding schemes, detailing how different configurations impact storage efficiency, redundancy, and recovery performance. Other critical considerations, such as network infrastructure, high-availability strategies, and workload-specific optimizations, are also covered to assist in designing a simplyblock deployment that meets both operational and business needs.

This guidance applies to all deployment topologies, with special sizing notes for hyper-converged Kubernetes and OpenShift deployments, where compute and storage share cluster nodes.