1 Welcome to Open OnDemand hosted by IFB cluster

1.1 About

The OnDemand portal serves as a unified, browser-based gateway to all your high-performance computing (HPC) resources within the IFB cluster.

Open OnDemand simplifies tasks by offering a graphical interface to view, edit, download, and upload files. Users can manage and create job templates for the IFB cluster, as well as access various interactive applications like Virtual Desktop, Jupyter Notebooks, RStudio and more. These functionalities are performed through a web browser, requiring minimal familiarity with Linux and scheduler commands.


Homepage

1.2 Highlights

Open OnDemand has a number of strengths that make it a user-friendly and adaptable solution for accessing HPC resources:

  1. User-friendly interface for easy access to HPC resources.
  2. Simple file and task management via a web-based interface.
  3. Access to interactive applications such as Virtual Desktop, Jupyter Notebooks, RStudio, etc.
  4. Reduced complexity, requiring minimal technical knowledge.
  5. Open-source platform fostering collaboration.
  6. Adaptability to diverse computing environments.

2 Getting started

To connect to Open OnDemand, visit https://ondemand.cluster.france-bioinformatique.fr

The first page of IFB Open OnDemand will bring you to a login prompt. Use your IFB cluster account to login.

Login

3 Features

Once you have logged into IFB Open OnDemand, the landing page will display several tabs across the top, including Files, Jobs, Clusters (shell access), Interactive Apps, and My Interactive Sessions. In the center of the page, direct links to popular interactive apps are prominently featured for easy access.

Homepage

4 Files

The File menu lets you view and use files in different file spaces: your home directory and the projects directory (/shared/projects). Selecting one of the file spaces takes you to a page that lets you navigate through the space in a graphical frame similar to a file browser on a personal computer. From here you can download, upload, create, delete and open files.

Files

The projects space contains all the “projects” directories hosted on the infrastructure. Use the text box filter to find your project quickly.


Project_filter

When you do not have access to a project, you will see an Error message like the one in the image below.

Project_acces_denied

5 Jobs

Jobs can be monitored, created, edited and scheduled with the job management tools under the Jobs menu.

5.1 “Active Jobs”

Under the “Active Jobs” tab you can view active jobs. You can choose to view your jobs or all jobs. From this menu you can also cancel your own jobs.

Active_jobs_all

5.2 “Job Composer”

Create and edit job scripts and schedule jobs under this menu.

Jobs_composer

6 Clusters (Shell Access)

The Clusters menu offers shell access to the login node within the IFB cluster. This shell terminal operates similarly to various other tools providing terminal access.

Upon initiating the Terminal, you will be prompted to enter your IFB cluster account password. Once entered successfully (no characters will appear on the screen as you type), your shell session on the login node will begin.

Shell_access

From here, you can work on the cluster using command line to run SLURM commands. If you are new to SLURM, please go through our documentations :

Please don’t run any bioinformatics tools directly on the login node. The login node is not a compute node ; it has no resources to run heavy calculation.

7 Interactive Apps

The Interactive Apps menu contains options to launch certain applications that have graphical user interfaces (GUIs) for interactive use on IFB cluster. Each application is running in a dedicated SLURM jobs with the resources (RAM, CPU, GPU) you request.

Current supported applications include :

  • JupyterLab : JupyterLab is a web-based interactive development environment for notebooks, code and data
  • RStudio Server : RStudio Server is web based version of the RStudio integrated development envrionment (IDE) for R

To find out more about using apps, please consult the following documents :

8 Starting a JupyterLab session

The JupyterLab environment lets you create and run Python, R and Bash notebooks and access a Unix terminal.

You can also install your own JupyterLab kernel with custom libraries. ?To check?

JupyterLab is the ideal environment for interactive data analysis, including machine learning.

  1. Select the “JupyterLab” choice from the Interactive apps menu.
  2. You can choose the cluster resources you need for your JupyterLab session. Such as project account, partition, number of CPUs, amount of memory.

Setup_JupyterLab_session

Click Launch to submit the JupyterLab job to the queue. The wait time depends on the current load of the cluster. Requesting smaller, shorter jobs may faciliate shorter wait times.

JupyterLab_starting

When JupyterLab is ready (running) click on “Connect to Jupyter”. Your JupyterLab session will be opened in a new tab.

JupyterLab_running

The JupyterLab launcher offers a collection of notebook editors, consoles and application like a Unix terminal, a diagram or text editor.

When you choose to create a new notebook by click a notebook icon, make sure that the file navigation panel located at the left is pointing to a directory where you have write access. Otherwise, JupyterLab will trigger an error message telling that you can’t create a new notebook file.

JupyterLab