Nimble Storage frees you from backup windows and the host performance impact required by traditional disk- and tape-based backup methods. Nimble’s VMware backup solution allows you to efficiently store months of VMware backups, while recovery takes only seconds.

Protecting VMware workloads with Nimble Storage is simple with the following tools and techniques.

Nimble Protection Manager for VMware

Making a consistent backup of virtual servers requires a file system quiesce to ensure that pending write activity is flushed to disk before the snapshot backup takes place. Applications such as databases also require a quiesce that asks the application to flush to disk any in-process database writes and pause the write process.

Nimble Protection Manager (NPM) is an efficient backup solution, designed to orchestrate the file system and application quiesce process with the snapshot point-in-time backup. NPM integrates with VMware to coordinate file system quiesce and Microsoft VSS APIs to coordinate application quiesce. This VMware backup functionality is included with Nimble Storage arrays and does not require separate purchase of host or guest agents.

Without Nimble, restoration of incremental backups can take hours or days.

Separate OS/Applications and Data

The storage characteristics of operating systems, application binaries and data are different and should be stored on different volumes to optimize performance, snapshot backup, and cloning. Operating systems and application binaries change infrequently enough that simple volume crash consistency is acceptable using any of the three methods for attaching storage to a VMware virtual machine.

Data volumes change constantly and benefit greatly from more frequent snapshot backup to improve recovery point objective. Separating them from the operating system also permits easy cloning for use in testing, reporting and data mining applications.

VMware Performance Policies

Nimble Storage allows you to create VMware storage volumes with specific Performance Policies that leverage application best practices for optimizing data storage and caching. User-definable block sizing is the key feature of Performance Policies that makes it a powerful VMware backup solution, optimizing storage block-size to match the production application’s logical write activity rather than writing them to larger general use fixed-blocks. This, along with real-time compression, greatly reduces the amount of data space required by snapshot-based backups which allows much longer retention than current approaches and significantly reduces the replication bandwidth required for off-site backup protection.

VMware Protection Templates

Protecting virtual machines has become a challenge as VM sprawl happens in most data centers. Nimble Storage provides VMware Protection Templates that provide virtual server protection rules that are uniformly applied across multiple volumes with similar protection requirements. Protection Templates consist of point-in-time snapshot retention policies, backup schedules, and replication schedules between arrays for local high availability and for off-site disaster recovery.

Volume Collections

Applications such as databases use multiple volumes that must be protected as a single unit to ensure application integrity. Volume Collections allow you to group these volumes to ensure that Nimble Protection Manager coordinates quiesce for all interrelated volumes. This functionality is especially important for database applications to ensure that transaction log and database volumes quiesce and snapshot are coordinated properly.

Nimble Storage Snapshots

Snapshot efficiency is of paramount importance to a VMware backup solution. Inefficient snapshot implementations constrain how frequently you can perform snapshots and how long you can retain those snapshots. There are two methods of performing snapshots, Copy on Write (COW) and Redirect on Write (ROW). When a write is performed on a data volume protected by a Copy on Write snapshot then the file system first reads and makes a copy of the block that is being changed to the snapshot storage. After the block is read and copied, then the write is allowed to change the block, file system indexes are updated and the system can continue processing other I/O requests. This methodology is very expensive in storage terms because it triples the I/O required for every write, thus reducing I/O performance by 66%.

Nimble Storage arrays use Redirect on Write snapshot methodology which avoids additional block copies and simply updates file system indexes. ROW snapshots dramatically improve storage performance by eliminating the additional I/O required by Copy on Write snapshots. And Nimble snapshots are compressed and stored on high-capacity SATA disks, further reducing costs.

To learn more about the Nimble Storage solution for VMware, click here.