hallta/distributed-sql

Distributed SQL

A way to configure a redundant SQLLite database across different nodes. Intended to federate across an array of Raspberry Pi devices, but should work across any hardware.

I just made this for fun. There's probably plenty of other well vetted solutions out there that do roughly the same thing. I didn't look.

Example Configuration

Below is an example of a four node setup. Below the diagram will explain what you should expect.

flowchart LR;
    RP1[ext.a] -->|firewall| fw(firewall);
    RP2[ext.b] -->|firewall| fw(firewall);
    app([App.a]) -->|RO| F;
    appb([App.b]) -->|RO| A;
    appc([App.c]) -->|RO| fw;
    fw --> A;
    A[main] --> B[(Storage.A)];
    A[main] --> C[Secondary #1];
    A[main] --> F[Secondary #3];
    C[Secondary #1] --> D[(Storage.B)];
    C[Secondary #1] --> E[Secondary #2];
    F[Secondary #3] --> G[(Storage.C)];
    E[Secondary #2] --> H[(Storage.D)];

Loading

1. Storage.{A,B,C,D} should all be eventually consistent.
2. main is the only write destination. This will ensure consistency.
3. App.{a,b,c} are all read-only apps. Can work from any node. Good for load distribution.
4. Setup as ext.{a,b} as public, with main behind a firewall. Good for external applications.

Setup

1. python3 -m venv env
2. source env/bin/activate
3. pip install -r requirements.txt
4. Confirm contents of config.json
5. $(which flask) --app main run --port <port>

See stats

1. Visit 127.0.0.1:<port> and dashboard should load