Using ActiveDR with Microsoft SQL Server Technology Overview

Everpure FlashArray

Built on all-flash storage, FlashArray provides storage and database administrators running SQL Server in their environments a fast, scalable, unified block- and file-storage platform that is ideal for high-performance SQL Server databases.

By providing a unified interface and simple-to-use tools for storage administrators, FlashArray gives those administrators the ability to quickly and seamlessly replicate, move, and manage data. FlashArray also deduplicates and compresses all data before it is written, efficiently reducing the size of data without impacting performance. Storage and database administrators can further increase storage by using the FlashArray snapshot capabilities to create snapshots of production databases, and they can use those snapshots in development or testing environments.

The FlashArray family consists of the following:

• FlashArray//C™: Provides low-latency storage for capacity-oriented workloads.

• FlashArray//X™: Provides high-performance, high-capacity storage that is ideal for performance-oriented workloads.

• FlashArray//XL™: Provides high-performance storage at scale that helps reduce the number of arrays needed to run large applications.

• FlashArray//E™: Provides economical-at-scale storage for workloads that aren't time-sensitive.

ActiveDR Overview

ActiveDR, which is included in Purity//FA 6.0 and higher with no additional licensing, provides storage replication to remote data centers that can help protect SQL Server data against threats such as hardware failures, ransomware attacks, and user error. This replication enables near-zero recovery point objectives and simple disaster recovery. As an integrated function, ActiveDR helps simplify disaster recovery workflows in situations where data must be continually protected. Disaster recovery workflows, including non-disruptive test failovers, live failovers, resync, and failback, can be easily configured and executed with minimal complexity. Specifically, test failovers with ActiveDR are non-disruptive to production instances, allowing organizations to validate disaster recovery readiness without interrupting ongoing operations. These test failovers can be performed without halting replication, ensuring continuous protection of production workloads while validating recovery processes in a seamless, transparent manner.

Latency-sensitive applications are not impacted by ActiveDR replication due to the feature's focus on front-end application performance. ActiveDR relies on asynchronous streaming replication, which is ideal for replication across longer distances with higher-latency wide area networks. Synchronous replication requires a target array to acknowledge every write to the source array. This type of replication requires low network latency between the source and target arrays. Asynchronous replication does not require a remote array to acknowledge application writes to the source array, which means that replication can occur across networks with higher latency. The asynchronous nature of ActiveDR means that storage administrators can use existing wide area networks to replicate data without concern for distance or latency.

ActiveDR consists of three components:

• Pods: These are storage-management containers that organize storage objects and configuration settings into groups that are failed over and failed back as a unit. A pod can contain volumes, volume snapshots, and protection groups. Additionally, a pod can contain configuration settings such as protection group snapshot schedules, snapshot retention policies, and quality of service volume limits.

• Replica Links: These provide replication between pods and provide directional and auto-reversing capabilities. Once a replica link is created, ActiveDR is automatically enabled.

• FlashArray Systems: ActiveDR requires a minimum of two FlashArray systems connected over a network to replicate data between. With no latency requirements, these systems can be within the same data center or as far apart as on different continents.

  

  FIGURE 2 | A typical ActiveDR deployment that includes pods and replica links between FlashArray systems �

Pod-based replication helps simplify storage management across sites by replicating all configuration changes made on a primary site array to the secondary site array, which helps simplify management and disaster recovery storage failovers. ActiveDR also supports multi-directional replication for different pods. For example, a database administrator might have a pod at their primary site that they want to replicate to their secondary site, while the secondary site might contain a pod that they want to replicate back to their primary site. ActiveDR lets them easily configure the pods to replicate in either direction between sites.

Microsoft SQL Server

SQL Server is a widely used relational database management system that has gained popularity among organizations of all sizes due to its scalability and ease of management. SQL Server provides high-availability features that database administrators can use, such as SQL Server Always On Failover Cluster Instances and Always On Availability Groups. These features can be enhanced using ActiveDR to provide database administrators with more disaster recovery options.

SQL Server also provides a number of features that make it an ideal platform for multiple workloads, from online transaction processing to complex online analytics processing. SQL Server handles structured data with ACID compliance that protects data and ensures reliable transaction processing. Data protection features, such as transparent data encryption and rolebased security, help organizations keep their data safe both at rest and in transit. As a Microsoft product, SQL Server offers deep integration with other Microsoft enterprise-grade technologies, such as Active Directory and Microsoft Azure, and it can seamlessly support real-time data analysis, reporting, and complex workflows.

Within this white paper, SQL Server is the core database platform that supports an organization's business applications and services.

Differences Between ActiveCluster, Asynchronous Replication, and ActiveDR

FlashArray has several storage replication functions that can be used to protect an organization's data. Among them are synchronous replication, which is used by the ActiveCluster™ solution; asynchronous replication storage array volume snapshots; and near synchronous replication with ActiveDR.

ActiveCluster

ActiveCluster uses synchronous replication to maintain copies of data between two FlashArrays. When data is written to a primary site FlashArray, it is simultaneously copied to a secondary site FlashArray. Once the data is written to both arrays, the write is acknowledged to the host system. This method of replication is recommended when the latency between arrays is 11ms or less, which means that ActiveCluster should be used between arrays in the same data center, or between data centers that have very-low-latency wide area network capabilities. For more information about ActiveCluster, see ActiveCluster Solution Overview.

Asynchronous Replication

Asynchronous replication is a snapshot-based solution that uses space-efficient snapshots to replicate data between FlashArray storage devices, while ActiveDR is a streaming-based solution that continuously replicates volume data between FlashArrays at different sites. When asynchronous replication is enabled on a volume at the primary site FlashArray, a snapshot of the volume is created on the primary site array and then replicated to the secondary site array. The first snapshot transfer is a baseline, which is a complete copy of the entire contents of the volume. All subsequent transfers are incremental transfers that result by comparing existing data on the storage array with the newly created snapshot to determine what data is sent to the secondary site array. For more information about asynchronous replication, see FlashArray Asynchronous Replication Configuration and Best Practices Guide.

ActiveDR

ActiveDR is different from the other replication solutions offered by FlashArray in that ActiveCluster is synchronous, with a zero recovery point objective and zero recovery time objective, whereas ActiveDR offers asynchronous replication, with recovery point objectives as low as five minutes. With ActiveDR, when data is written to a primary site array, the same data is immediately sent to the secondary site array. Worth mentioning that the data sent to the secondary pod is compressed, but it is not deduplicated before being sent. Deduplication occurs once the data lands on the secondary array. Because the data is sent asynchronously, the host that writes the data to the primary site array does not need an acknowledgement from the secondary site array. By using continuous replication, ActiveDR is an ideal solution for continuously replicating data between sites that are a greater distance from each other or that have higher-latency wide area networks between them. The continuous replication between sites can result in near-zero data loss, but when a failover is planned, SQL Server and the volumes protected by ActiveDR can be quiesced and replicated such that there is zero data loss.

For more information about ActiveDR, visit the Everpure support site