The 3-2-1 rule is a simple way to think about backups. Keep three copies of your data: the original plus two backups. Store them on at least two different kinds of media or devices, so one failure type can't take out both. Keep one copy offsite, away from the building where the original lives, so a fire, flood, or theft doesn't get everything.
In practice, a home setup might be your laptop, an external drive you back up to, and a cloud backup service. A small business might use a NAS in the office, a second copy on rotated external drives or tape, and an offsite or cloud copy. Newer advice adds that at least one copy should be offline or immutable, meaning ransomware or an administrator mistake cannot alter it.
Two things break backups more than the lack of a plan. One is never checking that they work: a backup you have not restored from is an assumption. Try restoring a few files, and a whole machine now and then. The other is syncing mistaken for backup. A folder that mirrors your computer to the cloud also mirrors deletions and encrypted files, so keep version history turned on.
Automate the process. A backup that depends on remembering will eventually be skipped.
Beginner · backup · prevention · cloud
SMART, which stands for Self-Monitoring, Analysis and Reporting Technology, is a feature built into hard drives and SSDs that tracks a set of health counters. Operating systems and free tools such as smartmontools and CrystalDiskInfo can read them.
Some attributes matter more than others. On a hard drive, reallocated sector count shows how many bad sectors the drive has already remapped. Current pending sector count shows sectors it wants to remap but can't read, and offline uncorrectable counts point the same way. Rising values in any of these are a warning to back up now. Spin retry count and command timeouts can point to motor or connection trouble. For SSDs, tools report wear indicators such as percentage used.
Be realistic about what SMART can do. A drive can fail suddenly with a perfectly clean report, and a drive with a few reallocated sectors can last years. A passing report is not a guarantee, and a warning is not a countdown. Its main value is catching a slow decline.
If SMART says the drive is failing, stop running repair tools and copy the data to another device first. Skip chkdsk and intensive scans on a drive with pending sectors.
Intermediate · smart · monitoring · prediction
A disaster recovery plan describes how a business gets its systems and data working again after something serious: a fire, flood, ransomware, a failed server room, or the loss of a cloud account. It is different from backup, which is just one ingredient.
Two ideas shape the plan. The recovery time objective is how long the business can be without a system before the damage becomes unacceptable. The recovery point objective is how much recent data it can afford to lose, which sets how often backups must run. A restaurant ordering system and an archive of old tax files have very different answers, and a plan sorts systems by importance so that the critical ones come back first.
A workable plan lists the systems and their dependencies, where backups are stored, who restores them, and how, including the passwords, license keys, and encryption keys that are needed. It also covers where staff will work if the office is unusable, and how customers and staff are told what is happening. Offsite and offline copies are essential, because a disaster that hits the building or the network should not reach every copy.
A plan that has never been tested is a guess. Practice restoring a system from scratch, time it, and write down what was missing. Update the plan whenever the systems change.
Intermediate · disaster · planning · continuity
A backup that has never been restored is a hope. Backup jobs can report success while the data they hold is incomplete, corrupt, encrypted by ransomware, or missing a folder nobody noticed was excluded. Testing is how you find out while it still matters.
There are levels of testing. The simplest is a spot check: restore a few files from different dates and open them. A better test restores a whole folder or database to a different location and checks that applications can use it. The strongest test rebuilds a full system, such as a server or a laptop, from the backup on clean hardware or in a virtual machine, and measures how long it takes. That number tells you whether your plan matches reality.
Many backup programs offer verification options that compare stored data with checksums. Those catch some problems, but not all, and they cannot tell you whether the right things are being backed up. Check the backup's scope against what the business actually depends on, including databases, email, configuration files, and anything on a laptop or in a cloud service.
Set a schedule for tests, and write down the result. Keep at least one backup copy offline or unchangeable, so that ransomware cannot encrypt it, and test that one too. When a test fails, treat it as the useful result it is.
Intermediate · backup · testing · verification
Time Machine backs up a Mac automatically to an external drive or a supported network location, keeping hourly backups for the past day, daily backups for the past month, and weekly backups until the disk fills. You can restore single files or an entire Mac.
It isn't a complete strategy on its own. A Time Machine drive kept next to the Mac shares its risks: theft, fire, power surges, and ransomware that can reach attached drives. Backups to network drives can also become corrupted after interruptions.
Check now and then that backups are completing, and add an offsite or cloud copy. If a Time Machine drive fails, it can be recovered like any other drive.
Beginner · time machine · mac · backup · restore
File History backs up personal files from your libraries and desktop to an external or network drive, keeping older versions so you can roll a document back. Windows also offers backup tied to your Microsoft account and OneDrive, and an older system image tool that copies the whole drive.
Each has gaps: File History doesn't cover programs or the system, OneDrive-based backup depends on sync and its retention rules, and the older system image tool is deprecated, though still present in many versions.
Whichever you use, confirm it's actually running, keep the backup drive disconnected or otherwise protected when not in use, and practice restoring a file now and then.
Beginner · file history · windows backup · system image · restore
Sync keeps the same files on every device, so a change anywhere happens everywhere. A backup is a separate copy with history, so you can go back to yesterday's version even if today's is damaged or deleted. An archive is data moved off active systems for long-term keeping, and often it's the only copy.
Problems start when one is mistaken for another: a synced folder treated as a backup, or an archive drive with no second copy.
Know which of your copies are which. Important data should have at least one real backup with version history, and archives deserve two copies on separate media.
Beginner · backup · sync · archive · versioning
Ransomware attackers look for backups first, because deleting or encrypting them forces payment. Backups on a drive that's always connected, or on a network share the attacker's account can reach, are exposed.
Offline backups are disconnected after each run, such as a rotated external drive or tape stored elsewhere. Immutable backups are stored where they can't be changed or deleted until a set retention period ends, using object-lock features in cloud storage or hardened backup repositories.
Use at least one of these for business data, protect backup administrator accounts with separate credentials and multi-factor authentication, and test restoring from the offline or immutable copy.
Intermediate · immutable backup · offline backup · ransomware · air gap
A surge protector diverts voltage spikes away from your equipment, and it should be replaced after absorbing a major surge; cheap power strips may offer little real protection. An uninterruptible power supply, or UPS, adds a battery that keeps devices running through short outages and gives them time to shut down cleanly.
NAS units, servers, and desktop computers benefit most, because sudden power loss can corrupt file systems and RAID arrays mid-write. Many UPS units can signal a NAS or computer to shut down automatically when the battery runs low.
UPS batteries wear out; test them periodically and replace them when the unit warns you.
Beginner · ups · surge protector · power · nas
Data on a drive in a drawer isn't permanently safe. Flash memory in SSDs, USB drives, and cards holds data as electrical charge that slowly leaks, faster when the cells are worn or stored warm. Hard drives hold magnetic data well, but their mechanics can stiffen or stick after long idle periods.
Store drives in a cool, dry place in anti-static bags, away from heat. Power them on and check the data periodically, and copy archives to fresh media every few years.
For anything important, keep two copies on different types of media, and don't rely on a single drive that hasn't been powered on in years.
Intermediate · long-term storage · shelf life · ssd retention · archive
When a RAID array, NAS, or server fails, the first questions are always the same: What RAID level? How many drives, in which order? Which controller and settings? Which drive failed first, and was any drive replaced? Was anything encrypted, and where are the keys?
Write it down while things are working: array layouts, drive serial numbers by slot, controller model and firmware, volume and share names, encryption and key locations, and backup schedules. Keep a copy somewhere other than the system it describes.
Good documentation doesn't just help a lab. It prevents panicked steps in the moment, which are a frequent way a recoverable failure gets worse.
Beginner · documentation · raid · configuration · business
On a modern encrypted phone, a backup is often the only way back after a failure. Turn on your phone's cloud backup, iCloud Backup on iPhone or Google's backup on Android, and confirm in settings that a backup completed recently. Photo sync, such as iCloud Photos or Google Photos, protects pictures separately.
Make sure the cloud plan has enough space, because a full account quietly stops backing up. For extra safety, copy photos to a computer or external drive now and then.
Also check messaging apps, authenticator apps, and notes apps, which may need their own backup settings.
Beginner · phone backup · icloud · google backup · photos
A NAS can tell you when something's wrong, but only if it's set up to. Configure email or app notifications for drive warnings, degraded arrays, failed scrubs, and temperature alerts, and make sure the messages reach someone who will act on them.
Schedule regular data scrubbing, which reads all data and checks it against parity or checksums, so problems are found while redundancy can still fix them. Review SMART reports for reallocated or pending sectors.
Replace drives that show warning signs before they fail, one at a time, with backups current. Don't wait for the array to go degraded.
Intermediate · nas · monitoring · data scrubbing · smart alerts