EyeFi Server 2 for Docker

Eye-Fi SD cards are designed to upload pictures and videos from your digital camera to a computer, using wifi. Initially these cards were supposed to work with the manufacturer's software, both on a local computer and on a cloud service.

Unfortunately the cloud service has been discontinued, then the manufacturer shut down; there is no longer an official support for the existing hardware.

Several projects filled the gap and allowed the cards to remain partially functional - See Attribution below.

EyeFi theory

EyeFi SD cards, when joining a wifi network, will scan computers on this network and connect to the first one listening on tcp/59278. This computer will be used as an agent to which the card will attempt to upload its contents.

Several SOAP exchanges will then take place between the card and the agent, over non-encrypted HTTP.

An authentication phase will ensure that the card will not upload its files to an unknown agent; this authentication implies that the agent know the card's wifi MAC address, and a shared secret known as the upload key.

Obtaining the EyeFi SD card MAC address and upload key

These values can be either obtained:

• From an existing and configured Windows installation of the EyeFi software (in the Settings.xml file under C:\Users\<user>\AppData\Roaming\Eye-FiX2 dir, or similar)
• From an existing and configured MacOS installation of the EyeFi software, using the following command: sqlite3 ~/Library/Application\ Support/Eyefi/Eyefi\ Mobi/offline.db 'SELECT o_mac_address, o_upload_key FROM o_devices'
• From Linux, using the excellent eyeconfig utility
• From MacOS, using an alternate build of the same eyeconfig utility

Docker image repository

This image is published in the Docker Hub repository: nikkoura/eyefiserver2

Using the docker image

Mandatory variables

This docker image is configured by environment variables.

• CARD_MAC: the wifi MAC address of the SD card. Digits can be separated by colons (:), dashes (-) or be specified without separators.
• CARD_UPLOAD_KEY: the upload key associated with the card.

Network

• The image port tcp/59278 must be exposed on the network where the card will connect.

Storage

• Uploaded pictures and videos are stored in a /uploads directory at the root of the docker image. A writable volume must be provided at this location.

Sample use with images published in Docker Hub

Using docker-compose

version: '3.7'

services:
  eyefiserver:
    image: nikkoura/eyefiserver2:1.0
    restart: "no"
    ports:
      - "59278:59278"
    volumes:
      - /mydata/uploads:/uploads:rw
    environment:
      CARD_MAC: "00:11:22:33:44:55"
      CARD_UPLOAD_KEY: "1234567890abcdef1234567890abcdef"

Using docker

docker run --detach -p 59278:59278 -v /mydata/uploads:/uploads:rw \
        -e CARD_MAC="00:11:22:33:44:55" -e CARD_UPLOAD_KEY="1234567890abcdef1234567890abcdef" \
        nikkoura/eyefiserver2:1.0

Sample use, building your own image

Using docker-compose

version: '3.7'

services:
  eyefiserver:
    build: .
    restart: "no"
    ports:
      - "59278:59278"
    volumes:
      - /mydata/uploads:/uploads:rw
    environment:
      CARD_MAC: "00:11:22:33:44:55"
      CARD_UPLOAD_KEY: "1234567890abcdef1234567890abcdef"

Using docker

docker build -t eyefiserver2 .
docker run --detach -p 59278:59278 -v /mydata/uploads:/uploads:rw \
        -e CARD_MAC="00:11:22:33:44:55" -e CARD_UPLOAD_KEY="1234567890abcdef1234567890abcdef" \
        eyefiserver2

Attribution

This is a fork of David Grant's EyeFI Server 2, which is itself based on the work of Jeff Tchang.