Sort guide

Cleaning an external drive without losing backups

An external drive is where your backups live — which is exactly why cleaning it is the riskiest cleanup you can do. The safe way treats the drive as a backup first and a cleanup target second. Here's how to free space without losing the one copy that matters.

Start with the 3-2-1 principle

Backup best practice is usually summarized as the "3-2-1" rule: keep at least three copies of your important data, on two different kinds of media, with at least one copy offsite. The exact numbers are a guideline, not a law — the point is redundancy. Your data should exist in more than one place, on more than one kind of storage, and at least one of those places should survive a single physical event like a fire, a flood, or a stolen laptop.

This matters before you clean an external drive because it changes what "safe to delete" means. A file on an external drive is not automatically a duplicate to remove — if that drive is your backup, it's your redundancy. Deleting it because it also exists on your Mac is fine only if you have a verified backup elsewhere. If the external drive is the only other copy, removing it leaves you with a single point of failure.

So the first question isn't "what can I delete?" It's "what would I still have if this drive died tomorrow?" If the answer is "nothing," the drive is not a cleanup target — it's a backup that needs protecting.

Plan first, delete never-by-default

The discipline that keeps you safe is to make a plan before anything moves. A plan is a list of what you intend to remove, where each file's verified copy lives, and why you're confident it's safe. It forces you to answer the "what would I still have" question for every file, one at a time, before any of them are touched.

This is the opposite of the "free space fast" instinct, and it's the right trade-off. A wrong bulk delete on a backup drive is not hours of work to undo — it can be permanent. The plan is what turns a risky cleanup into a deliberate one.

Concretely, a good plan names, for each candidate:

If you can't fill in all three for a file, it isn't a candidate yet. Leave it alone.

Verify before you delete

Before removing anything, confirm the copy you're keeping is genuinely identical. Equal names and equal sizes don't prove the bytes match. The reliable check is a checksum — a fingerprint of a file's bytes. Two files with the same checksum are byte-for-byte identical.

On a Mac, open Terminal and run:

shasum -a 256 "/Volumes/Backup/old/file.mov" "/Users/you/Documents/file.mov"

If both lines show the same hash, the files are identical and the copy you're keeping is real. If the hashes differ, they are not the same file — keep both.

For a large set of files, checksumming every pair can be slow. A reasonable middle ground is to sample: checksum the files that matter most, and spot-check a handful of others by opening them and confirming they're the right content. Sampling is honest only if you accept that it's a sample — it catches gross mistakes, not every subtle difference. For anything you can't afford to lose, checksum it fully.

Never clean the only copy — copy first, then clean

The single most important rule for an external drive: never clean the only copy of anything. If a file exists only on the external drive, it is not a duplicate — it's your only copy, and deleting it is data loss, not cleanup.

If you want to free space on a drive that holds the only copy of some files, the safe sequence is: copy first, then clean. Copy the files you want to keep to a verified new location, confirm the copy is good (checksums again), and only then treat the originals on the drive as removable. The drive becomes a cleanup target only after the data it holds exists safely elsewhere.

This is also why you should never clean a drive that is your only backup. If the external drive is the sole place your important data lives, the right move is to add a second backup before you remove anything — not to start deleting from the one copy you have.

Time Machine caveats

If the external drive is a Time Machine backup, cleaning it is a different and more delicate job. Time Machine manages its own backup structure, and it's not designed for you to delete individual files from it by hand. Removing files from a Time Machine backup can leave the backup in an inconsistent state, and the space you free may not be what you expect.

The honest guidance: don't hand-delete from a Time Machine backup. If you need space, the supported path is to let Time Machine manage its own retention — it prunes old backups over time — or to start a fresh backup on a larger drive. If you're cleaning a drive that is also a Time Machine destination, treat the Time Machine portion as off-limits and clean only the non-backup folders you know are yours.

And a plain caveat that applies to any external drive: drives fail. A drive that's been running for years, or one that's been dropped or shows errors, is not a safe place for your only copy of anything. If you're cleaning a drive because you're not sure it's healthy, the priority is copying your data off it, not freeing space on it.

When a tool actually helps

Manual planning and verification are the right call for a small, well-understood drive. They stop scaling when the drive holds thousands of files across many folders, and you can't realistically plan and checksum every pair by hand. That's the case where a tool earns its place — not to make the decision for you, but to surface the candidates and keep every step reversible.

Whatever tool you consider, hold it to the same standard as the manual workflow above. It should show you the full plan before anything moves, never delete by default, and give you a way back. If a tool deletes directly to free space fast, or asks you to trust its "smart" selection without showing you the plan, it's asking you to accept the risk all at once — and on a backup drive, that risk is permanent.

If your drive is small and you know exactly what's on it, you may not need any tool — the manual plan-and-verify workflow above is enough. A scripted finder exists for the cases where there are hundreds or thousands of near-identical files across multiple drives, where hand-reviewing every pair isn't realistic.

Where Archivist Sort fits

If the manual pass works for you, use it. When the pile is too big to review by hand, Archivist Sort applies the same careful workflow to the whole drive — with the safety rails kept in place:

Archivist Sort is designed around the same idea as the manual workflow above: inventory, review, approve, quarantine, rollback. It doesn't guess, and it doesn't delete by default. You still decide what's kept; it just makes reviewing thousands of files tractable.

It's built to back the claims it makes — see how it works on the product page.