Note: This is an automatically translated file. Original content from here:
This repository contains the necessary files to configure, build, and launch a Docker container, including docker-compose.yaml
, Dockerfile
, and scripts. Some environment variables are stored in a .env
file, which must be created with the create-env.sh
script so that some settings can be adopted by the host system. You also need a base Docker image that is automatically requested from Docker Hub.
The use of this repository is intended to help create Docker containers that provide the necessary requirements to be able to build flash images and packages with the Yocto/OE build system.
- 1. Requirements
- 2. Prepare
- 2.1. Clone repository and switch to the cloned repo
- 2.2. Configure environment variables
- 2.3 Volumes
- 2.4 Configure Ports
- [2.4.1 Web Access](#241-web access)
- 2.4.2 SSH
- 3. Build Container
- 4. Start container
- 5. stop container
- 6. Using the Container
- 6.1. Login
- 6.2. Use build environment
- 7. Update Container
- 8th. Support
Tested under Linux Debian 11.x, Ubuntu 20.04, 22.04. But it should work on any current distribution that runs Docker.
Necessary:
- Docker, Installation
- Docker Compose >= v2.24.3, Installation
- Git, installation via the package manager depending on the distribution
Optional:
- Docker Desktop
- Portainer, Installation (available with Docker desktop as a plugin)
Important! After you have installed Docker and want to run Docker as a non-root user, which is appropriate for our purposes, you should add yourself as a user to the "docker" group using this command:
sudo usermod -aG docker $USER
Then to apply the setting, either log out and log in again or restart!
git clone https://github.com/tuxbox-neutrino/docker-buildenv.git && cd docker-buildenv
Run this script to create the necessary .env
file:
./create-env.sh
The script gets some environment variables from the host system and puts them into an .env
file so that the container is configured to suit your host system. As a rule, that should be enough. Adjustments are usually not required. However, if this does not cover your requirements, you can adapt this generated .env
file. However, you should not run the script again, otherwise the .env
file will be overwritten again. It is therefore advisable to either rename this customized .env
file and rename it accordingly in the docker-compose.yml
file, or preferably to use another one in this form as a parameter when running docker-compose
- -env-file <my .env file>passed to
docker-compose`.
The container uses Docker Volumes to store persistent data, which allows access to specific files and directories permanently in the container.
In the standard configuration, these volumes are in principle integrated to suit the environment of your host system and mounted when the container is started, so that ideally you do not have to change anything in the volume configuration.
If you want to make changes to this, you can find the configuration of the volumes in docker-compose.yml
. Note that these settings are normally aligned with the paths as preconfigured for the Yocto/OE build environment using the init script from the Buildenv repository. If adjustments are made to this, you should take this into account!
These paths are mounted as volumes in the container. You have normal access to it via your host:
/home
└──<$USER>
├── tuxbox
│ ├── .config
│ ├── .data
│ ├── am
│ └── buildenv
├── Archives
├── am
├── sources
├── sstate cache
The container provides some access via certain network ports. This allows access to the build results via a web browser and access to the container via ssh.
By default, the container is configured to listen on port 80. Your host is mapped to the container's built-in web server (lighttpd
) via port 8080:
- Port: 8080 (host) -> 80 (container)
This enables access via web server to the generated images and packages (ipk's). Set-top boxes can therefore access updates directly from your home network, for example. If port 8080 on your host system is already occupied, you can either adjust these settings in the docker-compose.yml
file or specify them when starting the container. This could look like this if you map to port 8081:
- 8081:80
Settings on the web server can be made in the responsible lighttpd configuration file, which is available in the corresponding volume:
~/tuxbox
└──.config
└── lighttpd
└── lighttpd.conf
dir-listing
is activated in lighttpd.conf
, so that you can get by without additional content.
~/tuxbox/config/lighttpd$ cat lighttpd.conf
...
#server.compat-module-load = "disable"
server.modules += (
"mod_dirlisting",
"mod_staticfile",
)
dir-listing.activate = "enable"
Usually you access the container directly via docker exec
.
Since Git is already part of the container, an ssh server is also provided. By default, the ssh server is configured like this:
- Port: 222 (host) -> 22 (container)
- Password: = Username (as set in .env)
If port 222 is already occupied on your host system, you can adjust these settings in the docker-compose.yml
file, just like with the web server, or specify them when starting the container.
Login from the host system itself:
ssh $USER@localhost -p 222
Log in from another computer:
ssh <user>@<IP or hostname of the computer on which the container is running> -p 222
Run docker-compose wrapper:
./docker-compose build
Note: The preceding ./
must be taken into account here since the wrapper script is in the repo. The wrapper script calls docker-compose
as intended, but after automatically creating an .env
file as described in Step 2.2! This wrapper script takes all parameters common to docker-compose
. It only serves to reduce the effort required to enter commands to create the environment variables, which are provided via the generated .env
file.
Run docker-compose: with different .env file
Note: Hm repository includes an .env.sample
as an example. However, if desired, this must be adjusted and explicitly passed to docker-compose
when creating the container.
docker-compose --env-file <path to other .env file> build
docker-compose up -d
docker-compose down
You should know the name or container ID to log in. Run docker ps
to see which containers are currently available:
docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
9d6e0d280a9e tuxbox-build:latest "/usr/local/bin/star…" 41 minutes ago Up 41 minutes 0.0.0.0:8080->80/tcp tuxbox-build
For example, log in to the container with the Container ID
9d6e0d280a9e or the Container Name
like this:
docker exec -it --user $USER tuxbox-build bash
or:
docker exec -it --user $USER <CONTAINER ID> bash
You should see something like this prompt:
~/tuxbox/buildenv$
First you should make sure that the init script is up to date. Therefore, carry out an update:
~/tuxbox/buildenv$ git pull -r origin master
From now on you can use the build environment with the container.
After logging into the container, you are already in the buildenv
directory, where the init script is located. Now you can continue as described here.
The images and packages produced by the build system are made available via persistent volumes within your host home directory. By default, this location is intended for this:
/home
└──<$USER>
├── tuxbox
: ├── buildenv
: ├── dist
:
Note: If you have set up your volumes differently, this may of course differ.
The container provides a web server and can be accessed locally and on the LAN via port 8080 by default:
or on the LAN
As stated in Step 2.1, the repository that contains the recipe for the container can be updated regularly. To do this, go to the repository and run this command:
~/docker-buildenv$ git pull -r origin master
Then have the container created as described here.
For further questions, problems or support, open an Issue in GitHub or report in the [Forum](https://forum.tuxbox-neutrino.org /forum/viewforum.php?f=77).