Sign in to my dashboard Create an account
Menu

Keeping your critical SAN applications always on

Manish Thakur
168 views

Keep your san applications always on We all rely on emergency services and know that they must always be available without interruption. A small delay can make a huge difference in outcomes.

One of the highest priorities of IT teams is to ensure availability of critical applications. Loss of availability imposes enormous cost on the business, often causing lasting damage.

To ensure availability of these applications, you need a solution that not only mitigates the key risks to availability but also conforms to overall business requirements. Such a solution requires building redundancy at various levels. But one of the most critical aspects is to maintain continuous availability of the dataset required by the application. NetApp® ONTAP® software enables your infrastructure to serve all workloads, and it has a robust portfolio of data protection capabilities. NetApp SnapMirror® has been central to ONTAP data protection, because it lets you use NetApp Snapshot replication for backup and disaster recovery of individual volumes and storage virtual machines (SVMs).    multiple-LUNs-infographic Multiples LUNs in an application dataset with write order preserved. 

Protecting  SAN application datasets 

SnapMirror has evolved to provide synchronous replication for individual volumes starting with ONTAP 9.5. But ONTAP 9.9.1 (GA release) introduced SnapMirror Business Continuity, a SnapMirror variant that provides continuous availability for application datasets. This is a significant leap from providing synchronous replication—that is, zero recovery point objective (RPO) for individual volumes and logical unit numbers (LUNs). (Currently, this capability is available only for LUNs over FCP or iSCSI.) An application dataset could consist of multiple LUNs mapped to one or more volumes, which could be hosting different constituents of an application (for example, database, logs, redo logs). The replication of the dataset needs to maintain the write order across different LUNs of the dataset, which is critical to keeping the dataset consistent and enabling the application to function using the mirror copy.

To ensure continuous availability, you need to be able to immediately restore host access to the mirror copy when one of the copies becomes unavailable. A host transparently fails over to access the LUN from the mirror copy whenever a failure renders the primary copy unavailable. As a result, the end user doesn’t notice an impact when a disaster causes the host to access the mirror copy of the LUN. This capability also enables your storage team to plan sitewide maintenance operations on your systems without requiring an outage.

Get more than just continuous availability 

The two clusters that constitute the SnapMirror Business Continuity relationship must be either NetApp AFF or All SAN Array (ASA) clusters, but each can use a different AFF or ASA system—whether entry-level or high-end. This flexibility helps you design an optimal configuration for a specific requirement. And the mirror copy created through synchronous replication doesn’t need to sit idle, awaiting a crisis: You can use it for read access for application development and testing. You can also use a clone of the mirror copy for things like disaster recovery simulations without affecting protection.

Given the importance of application availability, there are several ways to accomplish it. Some applications have this capability built in, but SnapMirror Business Continuity provides the distinct advantage of maintaining availability for all critical applications—much simpler than using a different approach for each application. And manageability solutions for ONTAP such as NetApp Active IQ® Unified Manager make it easy to monitor and report on compliance to service-level objectives.

NetApp MetroCluster is another ONTAP solution that enables continuous availability of all workloads, creating a continuously available infrastructure. But MetroCluster doesn’t let you identify every individual workload that should be mirrored for availability. SnapMirror provides the granularity of control to identify datasets that need protection. Because SnapMirror functions on a standard ONTAP cluster with no special hardware required, it also inherits the cloud-ready DNA of ONTAP.

A familiar solution offering a head start 

SnapMirror is used by the vasta larger majority number of ONTAP customers. With so many storage administrators already familiar with ONTAP and SnapMirror, very little investment in learning—if any—is needed. Efficiency and powerreliability have been the hallmark of this solution, and it inherently supports the continuous stream of innovation in ONTAP. SnapMirror can be enabled on any system running ONTAP 9.9.1 and later releases with the Data Protection license installed.

You can learn more about deploying the solution here.

Manish Thakur

View all Posts by Manish Thakur

Next Steps

Drift chat loading