Docker containers

Automated Docker deployments¶

Docker containers provide a simple and efficient way to run software in an IoT or Linux environment. The Docker configuration bundle allows to automatically deploy Docker containers from any publicly or privately hosted repository.

Make sure Docker is installed on the system. The agent will use the docker version that is available on the device.

Once a defined docker container exits, it will be automatically restarted.

Definition of a container can be as simple as providing a name for it and an image to use to start it with. But it can be customized with optional arguments (passed to the docker command line when starting containers), an environmental file (which needs to be uploaded using file manager), as well as a command to run (which will override the default command from the provided docker image).

For privately hosted images, it's required to provide authentication credentials using the "Registry Authentication" tab. It requires a repository hostname to which we want to authenticate, as well as username and password. When using private repositories, the image name must include the repository hostname in it.

For example an image:

my-image:latest" -> hosted in "my.private-repository.com"

should be defined as

my.private-repository.com/my-image:latest

in a container definition.

In the device inventory tab it is possible to see additional information about the docker container.

If you need additional Docker functionality you might want to look at how qbee supports you with "docker compose" or jump right into a complex Docker compose use case deploying multiple Docker containers creating a full Node-RED, InfluxDB and Grafana system on an embedded edge device.