A backup you have never tested is just a hope
What you will learn
- The only backup rule you need to remember, in six words
- Why copying files is the easy half, and what the hard half is
- How my own backup turned out to be faithfully saving the damage
Everybody has backups right up until the first day they need one. That is when they discover the drive was unplugged in March, or that the folder they cared about was never included, or that the whole thing has been quietly saving nothing for months.
A backup you have never restored is not a backup. It is a wish.
Here is the part nobody wants to hear. Copying is the easy half. Any program can copy files. The hard half, the half that actually saves you, is proving you can get the files back. If you have never once done that, you do not have a backup system. You have a folder you feel good about.
Every drive you own is a clock ticking towards the day it stops, and it will not warn you first.
The only rule you need
Three copies of anything you would hate to lose. On two different kinds of storage. One of them somewhere else entirely.
That is it. The photographs that cannot be taken again and the documents that cannot be made again are the whole target, and they are smaller than you think.
Here is the part that hurts: the copy sitting next to the original does not count for much. If a power surge, a failed drive or a thief takes the machine, it takes the drive beside it too. The copy that saves you is the one that was somewhere else when the bad thing happened.
Now the part where I found out mine was lying
For a while my backup was a copying job to a second drive, running on a schedule and reporting success every night. It looked perfect.
Then I actually tried to pull back a file from three weeks earlier, and learned that it only ever kept the most recent copy.
Think about what that means. If something had corrupted or scrambled my files, the backup would have dutifully copied the damaged versions over the good ones, silently. It was faithfully backing up the damage.
The fix was a tool that keeps every version, plus a monthly habit I now refuse to skip: pick a file at random, restore it from last month, and open it.
If it opens, I have a backup. If it does not, I would rather find out on a boring Tuesday than on the worst day of the year.
One drive is a countdown. A backup you have never restored is not a backup, it is a wish.
A real backup for the files you cannot make again: kept in versions, scrambled so nobody else can read it, in three places, and tested. Replace anything in angle brackets with your own value.
1. Decide what you would actually cry over
What to do: not the whole disk. Only the things that cannot be downloaded again. Photographs, documents, keys, the settings for the things you built.
Why keep it small: a small correct backup beats an enormous one you never check. Write the list down.
<path-to-photos>
<path-to-documents>
<path-to-configs>2. Use something that keeps versions
What these do: the first sets up a storage area on your second drive that keeps every version and is scrambled so only you can read it. The second saves your chosen folders into it.
Why not a plain copy: keeping versions is exactly what saves you from a good file being overwritten by a damaged one and then backed up.
What should happen: the first run takes a while. Later ones are quick, because it only stores what changed.
restic -r <second-drive-path> init
restic -r <second-drive-path> backup <path-to-photos> <path-to-documents>Find it at restic.net, or borgbackup.org. Both are free, scrambled and keep versions
3. Get one copy out of the house
What this does: sends the same protected copy somewhere that is not your home. Cheap online storage, or a drive you keep at work or at a relative's and swap over.
Why it is safe to do this: because it was scrambled on your own machine in the previous step, whoever stores it only ever sees meaningless blocks.
restic -r <offsite-repo> backup <path-to-photos> <path-to-documents>4. Put it on a schedule
What this does: saves your files, then tidies up old versions so the storage does not grow forever. It keeps every day for two weeks, then every week for two months.
Why automate it: a backup you have to remember is a backup you will forget.
restic backup <paths> ; restic forget --keep-daily 14 --keep-weekly 8 --prune5. The step everybody skips: restore something
What this does: pulls one file out of an old version and puts it in a scratch folder so you can open it.
Why this is the only step that matters: it is the one that proves the other four worked. Everything above is a wish until you have watched a real file come back out.
How often: once a month, with a file picked at random.
restic -r <repo> restore <old-snapshot> \
--include <one-file> --target <scratch-dir>If it opens, you have a backup. If it does not, you found out on a good day.
This page in the original Kyber Cypher voice: Backups that actually work