I was an avid collector of .mod files back in the days of the Amiga, anyone remember those? There was a whole scene with floppy disks packed with samples, tracker programs with dense information-rich user interfaces. You could tweak a sample in a sound editor, swap out one sample for another, and generally mess with the music to no end. Getting to grips with the user interface was a steep learning curve.

Protracker 3.62 screenshot from Wikipedia

Protracker 3.62 screenshot, By Ivar Just Olsen, Bjarte Andressen, Tom Bech, Puw, Tom Beyer https://en.wikipedia.org/w/index.php?curid=57725881

At the time my Amiga hard disk was a whopping 105 Megabytes. Twice the standard size (that’s what she said, waka waka), and it still wasn’t enough.

A part-time job while I was at University gave me the opportunity to purchase a massive 1GB drive. One whole gigabyte! Amazing. This drive was substantially physically larger than my existing hard disk, and made a hell of a racket when running. You could tell when it was being read or written, no need for disk lights.

Old chunky Seagate hard disk

It wasn’t exactly like this but I remember it being chunky and contoured. Picture from https://www.alexandrugroza.ro/olddiscdrives

Fast forward a few years.

No, a few more.

Ok a lot more. Fast forward THIRTY years. To last night, in fact. I had finally started backing up my digital photos to cloud storage. Except somewhere in the depths of S3 configuration I had made a mistake and now the backup program couldn’t verify its work.

I spent a solid 30 minutes or more on the problem, ably assisted by Claude and Laphroaig1, but to no satisfaction. My existing backup bucket was a mishmash of intelligent tiering, standard storage and glacier deep archive. I had no patience to wait for all the glacier files to be defrosted so I could move them to intelligent tiering, I wanted immediate solutions.

For a while I hesitated. That photos folder was hundreds of gigabytes in size. Hundreds! For me, last night, that still seemed like a lot. I didn’t want to do the obvious thing - delete the entire backup and start again.

Then I realised. Size doesn’t matter! I wasn’t going to supervise every last file going into storage. It wasn’t going to clog up the telephone line or degrade my netflix streaming. Just do it lah.

It turns out the fastest thing you can do with Glacier Deep Archive files is delete them2. Millisecond response. 😆


  1. what a great combination, highly recommended ↩︎

  2. this might be the answer to an AWS interview question. I can neither confirm nor deny ↩︎