Where O Where is Your Backup Data and Can You Find It When You Need it?

All enterprise businesses backup their data but how confident are they that they can restore it? As more companies adopt disk as their primary target for backup, they can be lulled into a false sense of security thinking that their issues associated with recovery are gone. Nothing could be further from the truth. In fact, what some companies are finding out is that while their immediate backup problems are solved, their issues surrounding recovery are just beginning. Tape typically remains part of many companies' backup processes so the movement and recovery of data from tape media cannot be ignored and even recovering backup data from disk after it is copied to another disk system is not without its challenges.

A specific problem that companies need to keep in mind is how do they recover data after it is no longer kept only on the disk subsystem on which it was initially stored? Backup data can be moved from disk to tape or other disk systems by a variety of methods. The backup software may copy the backup data from disk to tape. The disk subsystem may replicate the backup job to another disk subsystem at another site. There are even some disk subsystems that can copy backup data to disk or tape. This all means that there are a lot of ways for backup data to get lost or misplaced after the backup to disk is complete.

Probably the best way currently for companies to avoid this scenario is to use the backup software to manage the placement and movement of all of the data short and long term to facilitate recoveries. Yet even in this situation, there is no guarantee that companies can recover data as rapidly as they might expect. The issue that companies run into is that as their production data stores grow, so do their backup data stores. As that occurs, the index or catalog that is needed to track how much backup data companies have and where it is at also increases in size.

The problem that then begins to emerge over time, especially when data is restored at sites other than the primary site or during disaster recoveries, is that companies need to first restore the backup software's catalog before they can restore any data at the remote site. How long this takes will depend on the backup software product in question as well as the size of the backup software catalog though I have heard stories that this catalog restore can take 24 hours or more to complete.

But not every product manages its backup software catalog this way. To mitigate this problem and track backup data regardless of where it is located or on what type of media, the CommVault® Simpana® software suite uses a distributed index. CommVault keeps a small MS SQL 2005 DB on the CommServer with pointers to a component of the index retained with each backup job. To recover the data, all that companies need is the CommVault Simpana server software installed at the recovery site, a copy of the small ER Backup of the SQL 2005 DB and the media (disk, tape or optical) with a copy of the Job on it.

The Simpana software then reads the data regardless of what kind of media it is stored on (disk, optical or tape) or how it got to the recovery site (Simpana may have vaulted the data or a disk subsystem may have replicated the backup job to another system at the remote site). In either case, Simpana can reconstruct the backup job and the data in it regardless of the type of media it resides based upon the information found in the index of each backup job.

The plethora of media types to which companies can store backup data and the way that data can be managed after it is backed up has mushroomed in the last few years. All of these options to store and move backup data are creating exciting new possibilities for driving down the ongoing costs of data management as well as opening new doors for data recoveries. But if companies forget to concern themselves with how to track and locate this data in an effective, efficient manner as they move it from media to media and site to site and not use a product that addresses these issues, they should not have a great deal of confidence that when it comes time to find and recover the data that they can recover it as quickly as the business may expect or need.