Storing data safely is becoming a big issue for home users, with many people now storing reasonably large amounts of important long-term data, such as digital photos. I've been asked a number of times recently about my views on the subject, so I thought I'd write some of this stuff down so I can point people at an article instead of reiterating everything verbally.
Many people have been storing their data on optical media such as CD-R. Problems with the longevity of CD-R media have been known for a while and more recently have received wider publicity. This is especially bad for data such as photos, which you likely won't refer to regularly so you won't notice the data going bad... until some time in the future (maybe 10 or 20 years) you come to look at your photos to find the whole lot irrecoverable.
Storing data on hard drives alone is similarly a bad idea - they are mechanical devices with associated wear and will die eventually.
RAID arrays help to reduce the risk of data loss from a bad disk, but
they should by no means be relied upon instead of backups - the power
supply in the machine can fail and fry all the hardware in the
machine (I've seen it happen, it's not a pretty sight). If that
happens when you have no backups, you've lost all your data in an
instant. A RAID also won't help you if some software (or user) goes
nuts and trashes your files - it only takes a small slip with the
rm command for you to lose your data.
In fact, probably the main use of a RAID array for storing sensitive data is to minimise disruption cause by a failure, rather than to help guarantee the integrity of that data. (There is obviously some benefit in the fact that in the event of a disk failure you still have your most current data, rather than losing everything since your last backup, but RAID should absolutely never be used instead of backups).
Minimising disruption is important in some cases - it is useful to be able to carry on doing your work after the disk has failed while you're waiting for the replacement to turn up, but this is probably not as important for most home-users wanting to store a bunch of photos safely.
Personally, I store my important data on my hard drive and make reasonably regular write-once full backups of the whole lot onto optical media (for me, that's about one backup a month, or sooner if I've made some big changes to my important data). The backup discs then get filed away safely and (hopefully) never touched again. If you're paranoid (sensible?), they can be stored off-site in case of fire, etc. too.
This method has some significant advantages:
rmcommand you can go back to a previous backup, even if you don't notice for 6 months, since you're never erasing the backup media.
Of course, there is also something to be said for making very regular
(nightly?) off-site backups to another server on the Internet using
rsync or similar differential file transfer software.
Especially if that server happens to be on another continent.
It's up to the individual how far they want to go to protect their data, but hopefully this article will give home-users a starting point when thinking about these problems. Remember that in many ways, a single digital copy of your photos isn't as robust as a printed copy. However, with proper precautions you can certainly mitigate those risks. Some risks affect the printed photos equally (i.e. fire, etc.), but it is far easier to mitigate those risks too when working with digital versions since it's so much easier to make backups of them.
Would you like to submit a comment about this article?