Beware of new storage technologies like disk-to-disk and inexpensive hard drives: they may offer extra space for data files, but they do not provide proper data backup and protection.
Data protection best practices require regular backups to tape, adherence to a proper tape rotation schedule, and off-site storage practices. So users requiring disk-disk near-line backup and restore capability must be sure to build in that last important step: disk-disk-tape that delivers a solid backup plan.
Backing up workstation hard drives and network servers is an unpopular chore, but easier methods and smarter technology can help. Two situations must occur for data to be lost: first the data is not backed up, and secondly the original copy is lost or corrupted. Human error accounts for one-third of all data loss occurrences.
Disaster recovery tends to be equated with serious data loss, but network administrators also find themselves responding to the loss of a laptop, a forgotten password on a ‘locked’ laptop, misplaced disks, or over-written files. And, occasionally, a crashed hard drive. According to InfoStor, problems arise from: natural disaster 3%, virus 6%, software or program corruption 14%, human error 33%, and hardware or system fault 44% of the time.
A significant issue is the inability to restore backed-up data. Such problems can be overcome by processes that prevent inadvertent over-writing of backup media and provide for good organisation of rotating backup libraries.
Continuous backup of all files is neither practical nor affordable. So opt for frequent backups of the most critical files that are revised and accessed often, and work towards less frequent backups of systems files. However, the sooner older data is archived, the smaller the backup data set will be.
So when do you back up? Develop a ‘backup window’ either when network demands are low, or by using an open file solution that allows backups while in use. Then implement a strategy that utilises partial, incremental, and differential backups as well as less frequent full system backups.
The following table shows the trade-offs between full and partial backups – essentially in time, cost and speed. To save on restore time and hassle in the event of data loss, choose a plan that includes full and differential backup. You can always perform a selective backup of critical files at any time. And don’t forget the data stored on desktop workstations and notebooks.
For differential backups, the volume of data (and time required for backup) increases throughout the week. Incremental backup requires fewer tapes and less time; however, consider the performance trade-offs. Two tape rotation methods are commonly used:
- Six-tape (or tape set) rotation-This comprises two (alternating) full backup tapes and one partial backup tape/day (based upon a five-day work week). Expanding to seven tapes gives a separate full backup for off-site storage, and avoids overwriting the only full backup copy.
- Grandfather-father-son scheme-This requires about 20 tapes (or tape sets for larger amounts of data). It utilises partial backups on a daily basis on the “son” tapes, full backup weekly on the “father” tapes, and full backup monthly on the “grandfather” tapes.
A key reason to consider business-class backup software is reliability when integrating full and incremental backups. Ultimately, the disaster recovery choice boils down to performance (and reliability), capacity and price.
A solid disaster recovery strategy is critical for any organisation, and is manageable when separated into its three core elements: backup, secure, restore. A balanced backup strategy helps protect critical, changing data frequently, while allowing appropriate backup of systems and setup files and expensive application programs:
- Data files = daily
- System and setup files = weekly
- Application programs = once