Guest Post: Protect Your Largest, Most Important Databases with Application-Aware Snapshots
In this guest post, 25-year data storage industry veteran Rich Vining discusses database protection and the importance of application-aware snapshots.
Many organizations, especially large enterprises, are facing a crisis in regard to protecting some of their largest and most important databases and applications from data loss or corruption. These applications need to be available to users and customers on a near-constant basis, and business line executives no longer accept that these applications need to be stopped to perform backups.
Indeed, as these databases have grown to multi-terabytes in size, the ability of traditional backup solutions to protect them has evaporated; it can often take more than 24 hours to perform a full database backup.
Some administrators are praying that their existing storage mirroring or replication solutions will save the day when something goes wrong, but this is dangerous because a database corruption or the deletion of important records will simply result in two bad copies of the database.
Array-based snapshot and cloning technologies offer a ray of hope. Both can perform a point-in-time backup in seconds, and more important, both can perform a recovery from a previous point-in-time copy in seconds. This effectively eliminates the backup window and drives recovery time objectives (RTO) down to near zero.
Also, because snapshots are so fast they enable administrators to perform backups more often, dramatically reducing the recovery point objective (RPO). Moving to one snapshot per hour, from a backup per day, reduces the amount of new data at risk by as much as 95 percent.
Crisis averted? Not quite.
Array-based snapshots and clones are not application-aware. This means that at the time a snapshot is taken, all components of recent database transactions may not have been flushed from the server’s memory to disk. The snapshot will capture only the parts that are on disk, and on recovery, you end up with an inconsistent (or corrupted) databases. Not good.
This is where application-aware snapshot management comes in to complete the solution. This software knows how to put the application into a backup-ready state, also called quiescing it. The software then calls the snapshot or clone function, and when that’s complete, it releases the application or database to start sending new transactions to disk.
Hitachi Data Systems arrays, such as the Hitachi Virtual Storage Platform (VSP) and Hitachi Unified Storage VM (HUS VM) offer high performance and highly scalable snapshot and cloning technologies: Hitachi Thin Image and Hitachi ShadowImage In-System Replication. They are integrated with Hitachi Data Protection Suite (HDPS), powered by Commvault, with IntelliSnap technology to perform intelligent, application-aware snapshot and clone management.
This integrated solution is the perfect answer to protect large, critical data sets in Microsoft Exchange, SQL Server, SharePoint Server and Hyper-V; Oracle and SAP; VMware vSphere; IBM DB2 UDB and Domino; and others. It helps increase storage efficiency and automates data backup and recovery for application, file systems and hypervisor platforms. The high-speed recovery of data can lower operating and capital expenses, reduce complexity and risk for Hitachi Data Systems storage customers.
Rich Vining is a 25-year veteran of the data storage industry, offering expertise and thought leadership in areas spanning document management on optical disks, archiving and data lifecycle management, data protection and unified recovery management.