Linux
Linux is available on request. The standard distribution is OpenSuSE 15.2 on desktop computers. Linux servers have SuSE Linux Enterprise Server or also OpenSuSE 15.2 (or higher). Linux will be installed as a Virtual Machine on top of MS Windows, meaning you will be able to run Linux and MS Windows at the same time, or as a stand-alone installation on your workstation. Dual-boot installations are not supported.
If you want Linux, please send in a request form. If you need support for Linux, email to helpdesk@mpi.nl with a mention of Linux. A lot of software is installed by default with some notable examples: Mozilla Firefox, Mozilla Thunderbird, Libreoffice, The Gimp, LaTeX, and development tools. The KDE desktop environment is the default installed window manager.
These is a linuxusers mailing list you can subscribe to in order to receive status updates and for general discussion.
Software on Linux can be installed through different methods
This includes the default operating system, GNU tools and many, much used, applications like Mozilla Firefox, Mozilla Thunderbird, Adobe Reader, The Gimp, LaTeX and development tools.
Users can't install these packages themselves and have to contact helpdesk when in need of specific software. These packages however can be updated by the user themselves when updates are available in the used repositories.
On request users can get sudo rights to install software themselves from the already configured repositories.
Software installed here is available for all Linux users when on a Linux workstation, a Linux virtual machine image or a selection of Linux servers running OpenSuSE 15.2 (lux20)
Software not in the path is installed in it's own sub-directory of '/usr/shared/apps' and can be selected by means of Environment Modules. With the command "module" you can get a list of available software and select what version you would prefer to be available in your personal environment. See here is a little howto on using the module command(pdf)
Contact helpdesk if you want certain software installed here to be available for other Linux users.
Systems running OpenSuSE 15.2 will support running software from singularity containers. The containers can be found here: "/usr/shared/singularity/containers".
The containers are the files with the ".sif" extension and are accompanied with a ".txt" file explaining the contents of, and the methods on how to use, the containers.
You can install and/or compile your own software in your own home directory or in the local "/usr/local" of your desktop on request. When binary executables or executable scripts are in "/home/<username>/bin" they will be in the default path and can be called from anywhere, the same applies for "/usr/local/bin". Binary executables or executable scripts installed in other locations in the home directory or a workspace have to be called with the absolute path in front of it or have to be added to your $PATH environment variable.
Compiling software on Linux might require a lot of software development dependencies, to ease the administration of every Linux installation these are not installed by default on every system. However they are installed on the Lux20 server. If you compile the wanted software on that server and install it in your home directory, the software will be available on all other systems as well.
- Perl
Perl is installed on every system as a part of the operating system as are some perl modules.
When other modules are needed they might be requested from helpdesk when they are available in the default software repositories of OpenSuSE. However a lot of modules are not available on the repositories and might be installed by the cpan command. Please contact the helpdesk if certain modules are needed.
To get a list of currently available modules type "instmodsh" and choose "l" for a listing
- Python
Python is installed on every system as a part of the operating system as are some python modules.
When a different version of python is needed it is possible to use a version of Python available in /usr/local by calling python with its absolute path or even define it as your default python by creating an alias to the alternative python executable in your .bashrc.
When other modules are needed you can contact helpdesk to make the requested module available for all in one of the Python installations in /usr/local or install it yourself in your home directory using one of the following methods on the Lux20 server (to have all development packages available):
easy_install --prefix=$HOME/local package_name
pip install --user package_name
To get a list of currently available modules type the following from the python shell:
help('modules')
- R
R is installed in /usr/shared/apps and can be loaded with the "module" command. If other versions or R are needed, or specialized R versions compiled against certain mathematical libraries as OpenBLAS or Atlas they might be available as different environment modules.
When certain R modules are needed you can contact helpdesk to make the requested module available for all in one of the R installations in /usr/local or install it yourself in you home directory using one of the following methods on the Lux20 server (to have all development packages available):
From within the R shell type:
install.packages(c("package_name"))
With installing in the personal library and choosing a repository nearby.
When you want to install a R package yourself but on a different location than the default in your home directory, for example to keep multiple versions separate or to use another file system like the cluster file system, use the following steps:
Create the folder you want to use:
mkdir /data/clusterfs/<department>/users/<user>/R_libs
From withing the R shell type:
install.packages("package_name", repos="http://cran.r-project.org", libs="<path_to_R_libs>")
To make sure R finds this folder every time when started create a file called ".Renviron" in your homedir with:
export R_LIBS=/data/clusterfs/<department>/users/<user>/R_libs
To get a list of currently installed modules type the following command from the R shell:
installed.packages()