MATLAB docker-compose.yml configuration

I'm trying to dockerize MATLAB with other services (React JS, Python Django, and Postgres database - these other services have been successfully dockerized and integrated).

When I run docker-compose up, it hangs on building the matlab image. I followed steps 1-4 (for step 2, I used the MATLAB Linux ISO image for MATLAB R2020b) for the Dockerfile for MATLAB: https://github.com/mathworks-ref-arch/matlab-dockerfile#optional-dependencies

I did not build the docker image (step 5) though as I need it to be built along with my other services. I noticed step 5 though requiresa a few parameters:

docker build -t matlab:r2020a --build-arg LICENSE_SERVER=27000@MyServerName .

What should the license server parameters be and how should I include them in the docker-compose.yml if necessary?

I copied the whole repository matlab-dockerfile afterwards to ./back-end/server_side/send_values/api/matlab/ (relative path on my MacOS system). The Dockerfile is in ./back-end/server_side/send_values/api/matlab/ and is the same Dockerfile at https://github.com/mathworks-ref-arch/matlab-dockerfile#optional-dependencies

How should I update the docker-compose.yml file to correctly build the MATLAB image?

docker-compose.yml

version: "3.9"

services:

db:

image: postgres

environment:

- POSTGRES_DB=postgres

- POSTGRES_USER=postgres

- POSTGRES_PASSWORD=postgres

matlab:

build: ./back-end/server_side/send_values/api/matlab/matlab-dockerfile

backend:

build: ./back-end

command: python server_side/manage.py runserver 0.0.0.0:8000

volumes:

- ./back-end:/code

environment:

- SECRET_KEY=${SECRET_KEY:?err}

- DEBUG=${DEBUG}

- DB_NAME=${DB_NAME}

- DB_USER=${DB_USER}

- DB_PASSWORD=${DB_PASSWORD}

- DB_HOST=${DB_HOST}

- DB_PORT=${DB_PORT}

ports:

- "8000:8000"

depends_on:

- db

- matlab

frontend:

build: ./react-ui

command: npm start

volumes:

- ./react-ui:/app

# Don't want to use local node modules

# https://stackoverflow.com/questions/29181032/add-a-volume-to-docker-but-exclude-a-sub-folder

- /app/node_modules/

ports:

- "3000:3000"

depends_on:

- backend

[Yash Patel]

Hey Aditya, were you able to get around this? I am trying to build an airflow data pipeline wherein I use a lot matlab for neuroimaging computations. Considering I am dealing with multiple containers i.e. Airflow and matlab I am currently orchestrating it with docker-compose but the issue I run into is that when i run docker-compose up it spins up the containers but matlab fails because it asks for a user email id and password which I can't provide as it is a multi-container setup. Further because I want to run neuroimaging stuff I need certain toolboxes which I hope you can help me figure out how to get them into this container. I'm sure I'm not first one having multi-container setups, is there a way to get the license sorted for this. I think my account has the necessary license to run this.

Answer 1

[Swastik Sarkar]

Hi @Aditya Behal, @Yash Patel,

It seems that you are attempting to build MATLAB as a service using Docker Compose. The method you are currently trying might require interaction if you do not have a network license server set up. To avoid this, I recommend using the "non-interactive" workflow with a MATLAB Batch Licensing Token.

This method is designed for scenarios where interaction is not feasible, such as in CI/CD pipelines or automated deployments. You can find more information about this method here:

https://github.com/mathworks-ref-arch/matlab-dockerfile/blob/main/alternates/non-interactive/MATLAB-BATCH.md#matlab-batch-licensing-token

Follow the instructions provided in the MATLAB Dockerfile repository to create a container image suited for non-interactive workflows. Detailed steps are available here:

https://github.com/mathworks-ref-arch/matlab-dockerfile/blob/main/alternates/non-interactive/README.md#create-a-matlab-container-image-for-non-interactive-workflows

Hope this helps.