When did you last actually test a restore?

enzosamaenzosama
in Technical

I keep telling myself my backups are fine because the jobs run and the logs look clean. Then I realised I have not actually restored anything in a long time, and a backup I have never restored is basically a hope, not a backup.

I am not asking which software everyone uses. I am curious about the habit itself. When was the last time you restored something for real, not just checked that a file exists in the backup? Did it work straight away, or did you find something missing, something out of date, or a step you had forgotten?

For a small setup with a couple of machines at home and a few rented ones, how often is realistic? Once a month sounds good on paper, but I suspect most of us only find out when we actually need it. What do you count as a test? Restoring one file, restoring a whole machine to a spare, or just reading the backup back?

My guess is a lot of us are in the same boat, jobs green, restore untested. Tell me when you last tested and what it found, if anything.

Comments

  • AnthonySmithAnthonySmith AdministratorProviderOG

    About a month ago, Debian + Nvidia is not a fun combination :)

    Time shift to the rescue

    TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA x2, USA x5, FR x3, AU, PL, NL, JP, BR, ZA
    FREE tokens on sign up, try before you buy. | Static Hosting Free for life: https://tierhive.com/static-hosting/

  • jadenjaden OG

    I try to spot check restoring backups once a month. Sometimes I feel lazy and don't do it, but I've been burned before, so I try to make it a habit.

  • david3david3

    I suppose I could automate test restoring a specific file and verifying it, but I guess I haven't felt the need. After setting a backup up, and then doing a test restore, has anybody found out later that it stopped working, despite the logs being good?

    I do sometimes actually restore from the backups for real. Because I accidentally deleted a file, or want to go back to an older version. Or because I'm moving a system. Just yesterday on my home system I wiped it and did a full restore of my system from a few weeks ago. Because I installed a bunch of libraries to do some specific builds that I no longer need and wanted to clean things up. And I don't have snapshots from that far back.

    It went OK. I remembered later that I wasn't backing up my home Downloads directory (intentionally). But I had been keeping some photos that I had selected and processed that I wanted to save. I still have the originals, but I'll have to redo the work.

    Also, my restore document was a little out of date, since I added another partition. So I updated the document.

  • yoursunnyyoursunny OG
    edited 9:20AM

    We performed a Seafile disaster recovery in 2026-Oct because Hostodo wiped our VPS during their maintenance.
    Nothing was missing, because we use the same procedure to move the Seafile instance between servers.
    The procedure is refined upon each migration / recovery, so that the next one runs more smoothly.

    The last flan is taken. 🍮 affbrr

  • rcy026rcy026

    I have a scheduled job that picks a file on random, locates it in the backup, restores it and compares it to the original. Cant remember that it has ever saved me from a corrupt backup though, it has worked every time.

    Last time I actively tested a restore was a few days ago. Very old Debian on a vps, managed to fubar something during upgrade so wiped it, installed newest Debian and restored from backup instead.

Sign In or Register to comment.