ActiveCluster

ActiveCluster
Audience
Public
Product
FlashArray
FlashArray > Purity//FA
Source Type
Documentation

Everpure® Purity ActiveCluster™ is a fully symmetric active-active bidirectional replication solution that provides synchronous replication provides RPO/RTO zero capability at the storage layer. It can be set up within or across multiple sites enabling clustered hosts and applications to be deployed into resilient active-active datacenter configurations.

Overview

ActiveCluster provides fully symmetric, bidirectional synchronous replication between two FlashArray systems. It is designed for resilient data center configurations where the same pod-based storage must stay available across arrays with zero RPO and zero RTO at the storage layer. It can be deployed within a site or across multiple sites, and it can also coexist with Active-Active asynchronous replication for broader business continuity designs.

At a high level, ActiveCluster uses pods as the unit of protection and consistency. A pod can exist locally on one array or be stretched across two arrays, and all objects inside the pod are managed as a consistency group.

Key Benefits

  • Zero-data-loss protection for stretched workloads because writes are acknowledged only after they are protected in NVRAM on both arrays.

  • Transparent failover behavior at the storage layer, which helps applications continue through FlashArray or site events when the host and application stack are designed for it.

  • Active-active access from either FlashArray, which supports metro-style clustered deployments and shared-storage application designs.

  • Simple building blocks for deployment: connect arrays, create a pod, place volumes in the pod, and stretch it.

  • Operational flexibility for planned migrations and future expansion, including the ability to pair with additional synchronous or asynchronous replication designs in supported releases.

Audience and Use Cases

This feature is for storage administrators, virtualization teams, platform architects, and application owners who need continuous availability across two FlashArray systems. It is especially relevant for VMware, Hyper-V, clustered databases, and other shared-storage-aware environments that can benefit from active-active storage access.

Use ActiveCluster when you need:

  • Metro-style business continuity with synchronous replication between two FlashArray systems.

  • Shared access to the same data set from either FlashArray for clustered or highly available applications.

  • Fast failover behavior with mediator-assisted arbitration and automatic recovery or resync after disruption.

  • A pod-based model that keeps a group of volumes consistent together.

How it Works

  1. You create a synchronous replication connection between two FlashArray systems by connecting them and selecting Sync Replication.

  2. You create a pod, then create new volumes inside it or move existing volumes into it.

  3. You stretch the pod to the second FlashArray, which performs an initial baseline transfer and then transitions to synchronous operation once the pod is online on both FlashArray systems.

  4. After the pod is synchronized, reads can stay local through preferred-array behavior, while writes are mirrored across both arrays before acknowledgment.

  5. If the replication path is disrupted, ActiveCluster uses recovery or mediator-assisted failover and then re-synchronizes the pod when connectivity returns.

Prerequisites

  • Two FlashArray systems are required.

  • For new IP-based deployments, the minimum required Purity//FA version is 5.2.4 or newer.

  • For new Fibre Channel-based deployments, the minimum required Purity//FA version is 6.1.0 or newer.

  • The two FlashArray systems can temporarily run different Purity//FA versions, but they must stay within two minor versions of each other, and matching versions are the best practice.

  • ActiveCluster requires four replication ports per array and resilient management connectivity so both arrays can reach the mediator.

For more information on the ActiveCluster implementation requirements, please see ActiveCluster Requirements and Best Practices.
Important: For the latest host, hypervisor, model, and feature combinations, use the FlashArray Compatibility Matrix and the Purity//FA Replication Requirements and Interoperability Matrix as the source of truth.

Deployment

At a high-level, to deploy ActiveCluster, you will need to perform the following steps:

  1. Connect the 2 FlashArray systems for replication.

  2. Create a pod on the first FlashArray with all of the objects that you wish to replicate.

  3. Stretch the pod to the second FlashArray and wait for it to become online on both FlashArray systems before connecting hosts on the remote side.

  4. After the base deployment, set preferred arrays, host personalities, and failover preferences as needed for your host topology.

For step-by-step instructions, go to the ActiveCluster Quick Start Guide.

Additional Information