docker.md

Docker

The Docker image bundles the Go server, Python fetcher, and frontend into a single minimal image (~150 MB). No nginx, no reverse proxy — the Go binary serves HTTP directly. Lance data is mounted from outside so you can point it at a local directory, NFS share, or S3 FUSE mount.

Quick Start

# Build the image
docker compose build

# Insert demo feeds + fetch articles
docker compose run --rm demo-data
docker compose run --rm fetcher-once

# Start server + continuous fetcher
docker compose up -d

Open http://localhost:8080.

Docker Compose Services

Service	Description
server	Go HTTP server on port 8080
fetcher	Python daemon that continuously polls feeds
fetcher-once	One-shot fetch (run with docker compose run --rm)
demo-data	Insert demo RSS feeds (run with docker compose run --rm)

All services share the same ./data volume. Replace it with any mount:

# docker-compose.override.yml - example: NFS mount
services:
  server:
    volumes:
      - nfs-data:/data
  fetcher:
    volumes:
      - nfs-data:/data

volumes:
  nfs-data:
    driver: local
    driver_opts:
      type: nfs
      o: addr=192.168.1.50,rw,nolock
      device: ":/srv/rss-lance/data"

Split Fetcher / Server Across Machines

Run the fetcher on one machine and the server on another, both pointing at the same Lance data:

# On your Linux server - fetcher only
docker run -d --name rss-fetcher \
  -v /mnt/shared-data:/data \
  rss-lance python fetcher/main.py

# On your laptop - server only
docker run -d --name rss-server \
  -p 8080:8080 \
  -v /mnt/shared-data:/data \
  rss-lance

Or without Docker - run the Go binary on Windows pointing at a Samba-mounted path:

net use Z: \\linux-box\rss-data
# Edit config.toml: path = "Z:\\"
.\build\rss-lance-server.exe -config config.toml

Build the Image Manually

docker build -t rss-lance .
docker run -d -p 8080:8080 -v ./data:/data --name rss-lance rss-lance