Analysis updated 2026-05-18
Experiment locally with database branching similar to PlanetScale.
Learn how Ceph snapshots and copy on write cloning work in practice.
Prototype a Kubernetes and Ceph based storage workflow before deploying it.
Test point in time restore and volume isolation on a single machine.
| homescale-dev/homescale | nelsonwerd/idea-to-ship-skills | ramset/pihole-ha-cluster | |
|---|---|---|---|
| Stars | 55 | 54 | 56 |
| Language | Shell | Shell | Shell |
| Setup difficulty | hard | moderate | moderate |
| Complexity | 4/5 | 3/5 | 3/5 |
| Audience | ops devops | vibe coder | ops devops |
Figures from each repo's GitHub metadata at analysis time.
Needs at least 20 GiB free storage, a Lima VM, and installs a full Kubernetes plus Ceph stack.
Homescale is an experiment in building something like PlanetScale, the commercial database branching service, but running entirely on your own computer. It lets you create writable database instances and point in time branches from immutable snapshots, without having to copy the entire database each time you want a new branch. Under the hood it combines a few well known open source infrastructure tools. Kubernetes acts as the control plane that manages everything. Rook operates Ceph, a distributed storage system, and Ceph RBD provides the block storage, the snapshots, and the copy on write clones that make cheap branching possible. All of this runs inside a single node Kubernetes cluster (via K3s) on a Lima virtual machine, so nothing needs to be installed directly on your host machine. To use it you install a tool called mise, then run a small set of task commands: one to bring the whole environment up (creating the VM, installing Kubernetes, snapshotting, Rook, and Ceph until the storage system reports healthy), one to test that snapshot restore and copy on write isolation actually work, one to gracefully shut the VM down while keeping its disks, and one to fully clean up and delete everything. The README is explicit that this setup has no redundancy or backup: there is only one monitor, one manager, one storage node, and one copy of every object. It is described plainly as an experiment, not a production ready system, and needs at least 20 GiB of free host storage to run.
An experimental local setup that mimics PlanetScale's database branching using Kubernetes and Ceph storage on your own machine.
Mainly Shell. The stack also includes Shell, Kubernetes, K3s.
The README does not state a license.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly ops devops.
This repo across BitVibe Labs
Verify against the repo before relying on details.