Create a docker container that contains a MATLAB install
No Data
You must perform these steps on a Linux platform. Before starting, you must install the following on the client platform - Docker - Git
The following steps guide you through the process of creating a Docker container image that contains a Linux environment with a MATLAB installation. Use the container image as a scalable and reproducible method to deploy MATLAB in a variety of situations including clouds and clusters.
Clone this repository to your Linux client using
git clone https://github.com/mathworks-ref-arch/matlab-dockerfile.git
Inside the cloned repository, create a subdirectory named
matlab-install
To install MATLAB into the container image, choose a MATLAB installation method. You can use MATLAB installation files or a MATLAB ISO image.
To obtain the installation files, you must be an administrator for the license linked with your MathWorks account. 1. From the MathWorks Downloads page, select the desired version of MATLAB. 2. Download the Installer for Linux. 3. Follow the steps at Download Products Without Installation. 4. Specify the location of the
matlab-installsubdirectory of the cloned repository as the path to the download folder. 5. Select the installation files for the Linux (64-bit) version of MATLAB. 6. Select the products you want to install in the container image. 7. Confirm your selections and complete the download.
matlab-installsubdirectory of the cloned repository.
installer_input.txtin the
matlab-installfolder. Move the copy up one directory level, into the root directory of the cloned repository.
matlab_installer_input.txt.
matlab_installer_input.txtin a text editor and edit the following sections:
fileInstallationKeyPaste your File Installation Key and uncomment the line.
agreeToLicenseSet the value to yes and uncomment the line.
product.MATLABto install MATLAB. Uncomment the corresponding line for each additional product you want to install. If you are not licensed to use a product, uncommenting the line does not install the product in the container. Your File installation Key identifies the products you can install.
Dockerfile. Edit the
Dockerfileand uncomment the corresponding line for each dependency you want to add. For more information, see Optional Dependencies.
Use the
docker buildcommand to build the image, using
.to specify this folder. Run the command from the root directory of the cloned repository. Use a command of the form:
docker build -t matlab:r2020a --build-arg [email protected] .Note: The
LICENSE_SERVERbuild argument is NOT used during the build but by supplying it here during build it gets incorporated into the container so that MATLAB in the container knows how to acquire a license when the container is run
To build a previous version of MATLAB, for example
Dockerfile.R2019b, use a command of the form
docker build -f Dockerfile.R2019b -t matlab:r2019b --build-arg MATLAB_RELEASE=R2019b --build-arg [email protected] .You must supply a tag for the image using the
-toption, for example,
matlab:r2020a. The tag names the repository for later use and deployment. Specify the location of the network licence manager using
--build-arg [email protected]. Replace
[email protected]with the port and location of your license manager. Alternatively, you can use a
license.dator
network.licfile to provide the location of the license manager. For more information, see Use a License File to Build Image.
For the R2019b Dockerfile you must also specify the MATLAB release using
--build-arg MATLAB_RELEASE=R20xxx, where
R20xxxrefers to a MATLAB release you are trying to build.
Use the
docker runcommand to run the container. Use a command of the form:
docker run -it --rm matlab:r2020a-
-itoption runs the container interactively. -
--rmoption automatically removes the container on exit.
Any extra arguments after the container tag are passed directly as command line arguments to the MATLAB process inside the container. For example, the following command prints
hello worldin MATLAB and then exits.
docker run -it --rm matlab:r2020a -r "disp('hello world');exit"
For some workflows and toolboxes, you must specify dependencies. You must do this if you want to do any of the following tasks. - Install extended localization support for MATLAB - Play media files from MATLAB - Run a network license manager inside the container - Generate code from Simulink - Use mex functions with gcc, g++, or gfortran - Use the MATLAB Engine API for C and Fortran - Use the Polyspace 32-bit tcc compiler
Edit the
Dockerfileand uncomment the relevant lines to install the dependencies.
If you have a
license.datfile from your license administrator, you can use this file to provide the location of the license manager for the container image. 1. Open the
license.datfile. Copy the
SERVERline into a new text file. 2. Beneath it, add
USE_SERVER. The file should now look something like this:
SERVER Server1 0123abcd0123 12345 USE_SERVER3. Save the new text file as
network.licin the root directory of the cloned repository. 4. Open the
Dockerfile, and comment the line
ENV MLM_LICENSE_FILE5. Uncomment the line
ADD network.lic /usr/local/MATLAB/$MATLAB_RELEASE/licenses/6. Run the docker build command without the
--build-arg [email protected]option. Use a command of the form
docker build -t matlab:r2020a .For more information about license files, see What are the differences between the license.lic, license.dat, network.lic, and license_info.xml license files?