An important element of any backup scheme is the retention of backups from various times in the past. For example, each week, you could place a backup tape into permanent storage. By doing so you are leaving a "trail" of backups that enable you to access data as it was at various times in the past. Here are some reasons for doing this:
Accidentally deleted file: You suddenly notice that a critical file is missing. It turns out that you accidentally deleted it three months ago while doing something else. Now you need it back.
Downgrading: A month after installing a new version of the operating system, you decide that it was a mistake and you want to go back to where you were before.
An added bonus of keeping a spread of old backups is that there will be lots of backups of a particular set of data. If, for some reason, the backups turn out to be unreliable (e.g. because of poor quality media) then at least you have several backups to try.
Failure to keep a spread of backups means that you are vulnerable to any data corruption that you don't detect immediately. For example, if you rotate your backups so fast that your oldest backup is just a week old, then any file that is corrupted and whose corruption is not detected for a week will be lost completely because all the backups will have been overwritten with the corrupted data.
It should be clear from this that frequent backups alone will not protect your data. In fact, a high turnover of backups combined with a short history of backups is more likely to result in the loss of data than is a slow turnover. So pay attention not only to the frequency of backups, but to their longevity too.
A later section on Integrity describes how to use the Veracity data integrity tool to protect against silent corruption. Maintenance of a history of backups, combined with regular use of Veracity provides very good protection against creeping silent corruptions.
Copyright © Ross Williams 1997. All rights reserved.