Node Configuration

Introduction 

The euroNAS HA cluster is designed to operate as a dual-node configuration, providing automatic failover in the event of a hardware failure. Nodes can be deployed locally (including within a single chassis) or across geographically separated locations, interconnected via WAN. This enables flexible designs aligned to both local resilience and site-level protection requirements.

The platform supports multiple operational models, allowing the cluster design to be aligned with specific workload and service requirements rather than enforcing a fixed architecture.

One model is based on independent storage per node with data synchronised through replication. In this configuration, failover occurs at the node level, maintaining service continuity with controlled data movement between systems.

An alternative model utilises euroNAS clustered drive technology, where both nodes operate against a shared storage layer. This enables continuous operation across the cluster, allowing failover to occur transparently without interruption to the application or requiring recovery actions at the workload level.

For deployments requiring scale beyond a dual-node architecture, multi-node capability can be achieved through the euroNAS eEKAS platform, which leverages Ceph to provide distributed storage and horizontal scalability across multiple nodes.

In this section, the cluster will be configured using a ZFS-based storage pool within a single location. 

NOTE: For deployments spanning locations over WAN, low round-trip latency is critical. A target of ≤12 ms RTT is recommended. When operating in synchronous mode, lower latency is strongly preferred, as write acknowledgement is dependent on remote commit. Elevated latency will directly impact write performance and may lead to application timeouts depending on workload sensitivity.