Definition
A backup strategy is the explicit plan that turns occasional file copies into a system you can actually recover from. It specifies what gets copied, how often, to what media, where the copies live, how long they are kept, and — the part most often skipped — how the recovery is tested. Without that last step a backup is only a hope; with it, a backup is a guarantee.
The canonical baseline is the 3-2-1 rule: keep three copies of your data, on two different storage media, with at least one copy off-site. The rule is deliberately simple because the failure modes it defends against are not exotic — a single drive dies, a single location burns, a single technology develops a bug. Independence is the whole point.
Why it matters
How it works
A full backup copies every selected file. An incremental backup copies only what has changed since the previous backup of any type, producing a small daily snapshot but requiring the full plus every incremental to restore. A differential backup copies everything that has changed since the last full backup, growing over time but needing only the full plus the latest differential at recovery. The trade-off is daily speed versus recovery complexity, and most production setups settle on weekly fulls with daily incrementals or differentials.
The technology choices on Linux range from low-level (tar piped to a file or network host), to file-aware (rsync, which copies only changed blocks), to image-level (dd, partclone). Each has its own failure modes. Whatever the tool, the strategy depends on three disciplines outside the tool itself: verifying integrity with checksums after each backup, keeping at least one copy on storage that is physically or logically disconnected from the production system, and performing periodic restore drills to catch silent corruption before you need the data.