July 16, 2026
Branch drama goes orbital
Let's Build PlanetScale from Scratch: Infrastructure
A DIY database dream sparks jokes, nitpicks, and serious side-eye from the crowd
TLDR: A developer is building “Homescale,” a DIY tool for making fast, writable database copies without duplicating everything. Commenters split between cheering the clever home project, cracking “MoonScale” jokes, and arguing the PlanetScale comparison is way too generous.
A developer’s plan to build “Homescale” — basically a homebrew spin on the idea of making easy, copy-free database branches — should have been a quiet builder story. Instead, the comments turned it into a mini tech soap opera. The project itself is simple in spirit: make it easy to create writable copies of a database without duplicating the whole thing, so developers can test changes fast and cheaply. It was inspired by old cloning tools, a failed past product, and, in a very online twist, memories of a Formula 1 job interview involving Ceph, a storage system.
But the real fireworks came from the peanut gallery. One camp loved the vibe immediately: “MoonScale” became the instant joke-brand, while another commenter said they’d already sent it to a friend at PlanetScale and were eagerly awaiting the reaction. That energy — half admiration, half mischievous workplace gossip — basically set the tone.
Then came the “hold on, that’s not what PlanetScale even is” crowd. The sharpest pushback argued that this project is being compared to the wrong thing entirely, saying it sounds less like PlanetScale and more like a standard managed database setup. In plain English: some readers think the name is overselling it. Others didn’t care and just wanted in, with one homelab tinkerer saying they were excited to try it and maybe contribute.
And in a delightful subplot, one commenter was less interested in the software than the F1 interview lore, begging for more details on what it’s like to work tech jobs in pro sports. So yes, the database project got attention — but the comments made it a full-blown fandom debate with bonus paddock gossip.
Key Points
- •The article presents Homescale as a project for creating writable database instances and point-in-time branches from immutable snapshots without full data copies.
- •Homescale uses a Docker-inspired model in which images are immutable starting points, containers are writable clones, and branches are writable containers created from an existing container state.
- •The initial examples use Postgres, but the storage model is intended to be database agnostic for engines whose durable state lives on a filesystem backed by a block device.
- •Engine-specific logic is meant to sit behind an adapter, while Homescale manages lifecycle concerns such as volumes, immutable states, writable clones, workloads, and lineage.
- •The design requires branching to occur in the storage layer so unchanged data can be shared, and the article compares this approach with architectures used by Amazon RDS, Google Cloud SQL, Aurora, AlloyDB, and Azure SQL Hyperscale.