This repo is a fork of this excellent repo https://github.com/cptactionhank/docker-netatalk to updating and improving the base implementation.
- Changed Debian version from Jessie to Buster-slim
- Install netatalk from official Debian repository
- Added
docker-compose.yml
for ease of use - Option to create multiple users
- Updated afp.conf to remove anonymous access
- Remove avahi ENV var (always run)
- Other configurations by env vars
To quickly get started with running an Netatalk container first you can run the following command:
docker run \
--detach \
--publish 548:548 \
manuellr/netatalk
Important: This does not announce the AFP service on the network; connecting to the server should be performed by Finder's Go -> Connect Server (CMD+K)
and then typing afp://[docker_host]
.
Default configuration of Netatalk has two share called Share which shares the containers /media/share
and called TimeMachine which shares the containers /media/timemachine
mounting point. Host mounting a volume to this path will be the quickest way to start sharing files on your network.
docker run \
--detach \
--volume [host_path]:/media/share \
--volume [host_path]:/media/timemachine \
--publish 548:548 \
manuellr/netatalk
Alternatively, it can be executed using docker-compose with the following command:
docker-compose up -d
With the slower roads documentation some knowledge in administering Docker and Netatalk assumed.
There are two ways of configuring the Netatalk which is either by mounting a configuration file or editing the file from the container itself. Documentation of the configuration file /etc/afp.conf
can be found here.
This is quite a simple way to change the configuration by supplying an additional docker flag when creating the container.
docker run \
--detach \
--volume [host_path]:/etc/afp.conf \
--volume [host_path]:/media/share \
--volume [host_path]:/media/timemachine \
--publish 548:548 \
manuellr/netatalk
That variables could be setted in a file called netatalk.env
. This file is used by the docker-compose.yml file or by using the flag --env-file
when exec the docker run
command.
Variable | Description |
---|---|
AFP_SPOTLIGHT | (yes/no) Enables the possibility that, when searching in spotlight on the mac, shows results of these volumes |
AFP_ZEROCONF | (yes/no) Enables the possibility that it can be detected on local networks. This must be complemented using host network when executing the container |
AFP_NAME | (def: Netatalk-server) Name of the device to be displayed |
To setup access credentials you should supply the following environment variables from the table below.
Variable | Description |
---|---|
AFP_USER | create a user in the container and allow it access to /media/share |
AFP_USER_PASSWORD | password |
AFP_USER_UID | (optional) uid of the created user |
AFP_USER_GID | (optional) gid of the created user |
If you need to add more user you can do it adding the user number. For example, AFP_USER_2
, AFP_USER_2_PASSWORD
or AFP_USER_15
.
This image includes an avahi daemon which makes it discoverable on the network. Enable by setting the environment variable AFP_ZEROCONF=true
.
Service discovery works only when the container use the same network as the users which is why you need to run the container in host
network, but do consider that it is considered a security threat. That option could be enable be supply --net=host
flag to Docker or by enabling docker-compose.override-net.yml
with the command:
ln -s docker-compose.override-net.yml docker-compose.override.yml
Alternatively you can install and setup an mDNS server on the host and have this describing the AFP service for your container.
Thanks to @rrva for his work updating this image to Netatalk version 3.1.8 and slimming down this image for everyone to enjoy.
This image has been created with the best intentions and an expert understanding of docker, but it should not be expected to be flawless. Should you be in the position to do so, I request that you help support this repository with best-practices and other additions.
If you see out of date documentation, lack of tests, etc., you can help out by either
- creating an issue and opening a discussion, or
- sending a pull request with modifications
This work is made possible with the great services from Docker and GitHub.