Processor Magazine — November 6, 2009
The concept of storage backups is simple: Copies of your enterprise’s data are created and made readily available in case they are needed. But as admins know all too well, putting backup plans into practice can be fraught with trouble.
The list of what can go wrong is long: Files on storage media can become corrupted, tape could get damaged, or software might not work like it is supposed to. Admins down in the trenches generally have many anecdotes to tell about the pitfalls associated with backups. To help make storage backup easier, here are some solutions to a list of common problems you are likely to face.
Data That’s Not Prioritized
Those in charge of backups often only concentrate on making sure that the source data is copied and backed up somewhere, with little attention given to which data is crucial and which is less important.
This inattention can be attributed to what Haim Sternberg, president and owner of Cherry Systems, says is a lack of effective policy before backup systems are put into place.
“The typical error before the backup is run is the implementation of the ‘policy’ that defines what, how often, file selection method (full, incremental, differential), and destination (tapes, backup servers, etc),” Sternberg says. “The backup operator then follows the policies, and the backup software has no specific knowledge of the relative importance of files among the thousand or more files it is about to back up.”
Because all the files are of equal importance, the admin picks the volumes and folder that appear “important” based on the policies. The sad result is that critical files do not get backed up, Sternberg says. “Missing among others [may be] the folder named Temp_pr_process, which is the folder the payroll department created to prepare the batch file for processing,” Sternberg says. “Of course, there are also infinite variations of this tale.”
The remedy is to establish which files are those that absolutely must be copied, compared to the rest of the data. “You need to focus on the full 100 files that cannot be lost, but instead one often ends up protecting everything,” Sternberg says. “You need to communicate what is important so the crucial files get backed up.”
It is crucial to think about what is really important from both a business and technology perspective, as well. Restored data, for example, must meet the enterprise’s essential business needs. Recovered data must also be able to support the systems that have to be running so the enterprise can continue to function, says Tom Cornwell, storage and virtualization architect for Hosted Solutions (www.hostedsolutions.com). “Another way to look at it is to determine what you absolutely cannot afford to lose,” Cornwell says. “It’s about what the consequences are by not having the data.”
Backups Interfering With Workflow
Users outside of the IT network appreciate backups when they are necessary but will object when the process interferes with their workflows. Although some admins may take offense to users complaining about backup systems given their importance, the task of copying data and applications should, in fact, be transparent to the end user, says Jeff Moyer, director of worldwide storage services at HP Enterprise Services.
“Today’s mobile workforces demand an efficient, network-based backup capability. The backup software and processes to maintain data must be nonintrusive to the normal use of the employee’s workstation,” Moyer says. “In addition, organizations need to think about and provide a way to securely transport and store data while providing user-friendly options to allow user-driven restore of data.”
However, users need to meet the IT department halfway by not doing anything that will hinder data backups, either. “It is important to keep users from removing files from the backup logic that prevent an effective backup from being taken,” Moyer says.
No Monitoring
Maybe you have installed a system that sends data offsite three or four times per hour so that your enterprise’s data is continually mirrored. In the event of a massive-scale system failure or even if your enterprise itself is destroyed in a fire or another disaster, it is possible to have all of the data and applications restored in a matter of minutes. But after initially making sure that the backup system is up and running, have you been monitoring things to make sure the system is doing what it is supposed to do? Many admins do not, says John Matzek, co-chief executive officer of Logic IT Consulting.
“It is fairly common where a backup may fail and nobody realizes it,” Matzek says. “Sometimes backups will not work for a long time and nobody catches the problem until there’s a disaster.” Therefore, it is important to keep tabs on what is going on by using monitoring software, Matzek says.
Incomplete Backups
It happens a lot: Files are open or in use and thus do not get backed up. “Another issue is to ensure that all of the files are included in the backup,” HP’s Moyer says. “Many times, databases or applications have open files that backup software bypasses if they are in use.”
The solution is to make sure that your software is able to overcome this all-too-common glitch. “It is important to create a schedule that shuts down applications or databases on a regular basis for ‘cold’ backups,” Moyer says. “Additionally, frequent backup transaction logs will allow for transactional recovery.”
Key Points
© 2007-2010 Hosted Solutions, LLC All Rights Reserved | privacy