A little while ago I blogged about how SnapVault and Deduplication work together for efficient D2D backups. SnapVault allows two NetApp systems to back each other up using block-level incremental snapshots and automatic deduplication. But what if you think that SnapVault is so cool you want to use it for all you backups, including DAS data stored directly on servers? In this blog I’ll tell you how you can do this with another type of SnapVault - Open Systems SnapVault, otherwise known as OSSV.
To describe the difference between SnapVault and OSSV in simple terms, SnapVault software resides in the NetApp storage controller while the OSSV software resides in an application server. The server being backed up can either be in the data center or in a remote location.
Some new features we included in OSSV 3.0 are:
- Support for Microsoft SQL Server
- Support for Microsoft Cluster Service
- Dynamic throttling
- Filter driver
- IPv6 support
- Support for volume mount points
- New platform support – Win2008 R2, Solaris 10, AIX 6.1
- Link compression
Now comes the clever part. OSSV operates by making an application server look like a NetApp primary storage system. Here’s how it works:
First, an OSSV software agent is installed on the application server you want to backup.
Next a baseline, or full, backup of the server is sent to the NetApp system. During this baseline transfer, the OSSV agent keeps track of a) all file modification times and b) checksums for each 4K block of data sent to the NetApp system. Once the baseline copy is made, all subsequent backups are block-level incremental.
This brings us to the final and cleverest step of all. For its incremental backups, OSSV first looks at file modification times. If the mod time of any file is the same as the last time it was backed up, OSSV doesn’t send the file (it hasn’t changed!). If the mod time has changed, OSSV runs a block checksum routine on the file and only sends the blocks that have changed.
OSSV can be run via command line scripts, through the NetApp Protection Manager console, or via standard GUI’s from CommVault, Syncsort or Backbone.
Oh, I almost forgot about deduplication. As with regular SnapVault, if dedupe is enabled on the NetApp OSSV destination volume, it will automatically sweep through all the 4K data blocks, removing any duplicate blocks that might have snuck through. That's pretty clever too.
My blog has been a very brief overview of OSSV just to whet your appetite. For more detailed information on OSSV technology and best practices I highly recommend reading this technical report.