Reusing spare computers for a personal server cluster

I arrived at the dilemma last summer of around thirty unused computers sitting in my garage, clouding both my car and my thoughts. Noticing also a conspicuously empty stretch of shelving in my closet, I elected to set up my own server room, constructed only from 8 year old Dell Optiplexes. After endeavoring to remove those that were faulty or otherwise malfunctioning, I ended with a selection of seven, and an old network switch. Armed only with a limited selection of linux experience and little knowledge of how servers even work, I resolved that within a week I would have a fully functional distributed computing system: accessible from anywhere, and free from Microsoft Azure’s grasp. Unfortunately, my first days did not go as planned, as I had to reinstall Debian upwards of three times on the first machine before it worked as planned. After repeating this process on each of the machines, I ended up with a horde of SSH accessible servers, and no idea where to go. The next couple hours were spent furiously googling–it couldn’t be that hard to just install kubernetes. Unfortunately, it was, and the jargon-laden articles about swarms and docker were just beyond my grasp. Somewhat overwhelmed by the numerous different recommendations, I decided to just run apt-install and see what happened. Over the next several days, I took repeated notes on the behavior I was seeing, and slowly got one server connected to a cluster instance, then another, eventually reaching a quorum with seven networked pcs. Unfortunately, my triumph soon became dismay, as I didn’t actually have any apps to deploy, or really any reason to need servers. Sitting alone at my laptop and staring at an unutilised resource list, I didn’t even care about using the servers, I just wanted to build them.