EUDAQ dockerfile

Creates the environment to run the EUDAQ framework. This image is based on an phusion/baseimage built over a ubuntu-18.04, and contains the necessary packages to run (or develop) the EUDAQ framework in a real test-beam setup. In order to use LCIO and EUTelescope with this image, take a look at dockerfiles-eutelescope package.

Installation

Assuming git, docker and docker-compose is installed on your system (host-computer).

Clone the docker eudaq repository and configure it

$ git clone https://github.com/duartej/dockerfiles-eudaqv1
$ cd dockerfiles-eudaqv1
$ source setup.sh

The setup.sh script will create some docker-compose*.yml files. It also creates the directories $HOME/eudaq_data/logs and $HOME/eudaq_data/data, where logs and raw data will be sent in the host computer.

Note that a eudaq repository will be cloned at $HOME/repos/eudaq (unless it is already in that location). This repository will be linked to the containers in development mode.

Download the automated build from the dockerhub:

$ docker pull duartej/eudaqv1:latest

or alternativelly you can build an image from the Dockerfile

# Using docker
$ docker build github.com/duartej/eudaqv1:latest
# Using docker-compose within the repo directory
$ docker-compose build eudaqv1

Usage: production environment

The production environment uses the EUDAQ v1.x-dev branch.

The recommended way to launch all needed services is with docker-compose You should be at the dockerfiles-eudaqv1 repository folder and launch:

$ docker-compose -f docker-compose.yml -f production.yml up

One service per each element of the framework (run control, logger, data collector, online monitor, TLU producer, ... <to be defined which are the minimum needed>) is created, all connected to the run control at tcp://172.20.168.2

To run only one particular service:

$ docker-compose -f docker-compose.yml -f production.yml run --rm <service_name>

note, however, that run control and the logger are always launched as needed for any of the EUDAQ producers or components. service_name could be:

runControl
logger
dataCollector
onlineMon
TestProducer
NIProducer
TLU

If you want to add other element of the framework, just create a container using the duartej/eudaqv1 image. Be sure you connect the service to the <foldername>_static_network (check your available networks docker network ls); and assign an unused ip:

$ docker run --rm -i \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    -e DISPLAY=unix${DISPLAY} \
    --network <foldername>_static_network \
    --ip 172.20.168.XX \
    duartej/eudaqv1
# Once inside the container, lauch the process you are 
# interested, and remember to connect (if needed) to run control
# at tcp://172.20.168.2

DISCLAIMER: in alpha yet, not tested in production environments

Usage: development environment

The development environment uses the EUDAQ repository placed in the host computer at $HOME/repos/eudaq, which was previously cloned and checkout to v1.x-dev branch in the installation step.

Analogously to the production environment, the recommended way to launch all needed services is with docker-compose, this time without explicitely especify the yaml files (as uses the default and the override mechanism).

$ docker-compose up

or launching a concrete service as explained in the production section:

$ docker-compose run --rm <service_name>

An extra service is available in order to allow compilation of the developed code: devcode and devcode-p (the privileged version, to be run for TLU related check). The build directory in the container is found in the /eudaq/eudaq/build:

$ docker-compose run --rm devcode