Overview#

There are a few points you should be aware of when working with Jupyter Notebooks on the jupyterhub server:

  • Sessions are persistent. If you close the browser tab without stopping the server before, you can revisit the jupyterhub site and carry on working

  • Closing browser’s tab or log-in out will not stop your session, your session will continue consuming resources (see next point)

  • It is preferable to stop the session when you finish. This depends on the jupyter interface:
    • JupyterLab: File -> Hub Control Panel then stop.

    • Classical Jupyter: Control Panel and then stop

  • Currently, in container mode we extended the top right bar with a Hub Control Panel button.

../../../_images/hubcontrol.png

Note

If you are new to DKRZ and want to access our JupyterHub, please consult the Quick Start page.

Note

If the batch queue is filled with waiting jobs, it might take some time to start the notebook server. There is an internal timelimit (300 sec) specified until which the resources should be ready. If not the spawner will reject and ask you to start again (maybe with reduced resources specified).

Introductory presentation#

On September 1st, 2020, we deployed the first customized release of our Jupyterhub server. It was presented in the DKRZ Tech Talks. You can view the talk on Youtube or download the slides. Since the initial release, we’ve made numerous improvements to the service that can be found in the Changelog.

Architecture#

../../../_images/jhub-arch.svg

Interface#

Jupyterlab is the default interface. However, you can switch to the classic notebook by adding /lab before /tree in the URL. To return to JupyterLab, simply reverse this change.

/user/<username>/<server-name>/lab/tree/ -> /user/<username>/<server-name>/tree/

If you use the advanced spawner, you can select the interface from the list of UIs before starting the session.

Issues with Jupyterhub?#

If you have issues while spawning servers or when working on your notebook, we suggest that you first try one of the following steps:

  • JupyterHub provides clear error messages for invalid inputs, access denials, and issues related to SLURM (JobID and REASON).

../../../_images/acces_denied.png
../../../_images/slurm-status.png
  • Check the log file if you see a waiting to connect mesage. By default, stdout and stderr of the SLURM batch job that is spawned by Jupyterhub is written to your HOME directory on the HPC system. In order to make it simple to locate the log file:

    • if you use the preset options form: the log file is named jupyterhub_slurmspawner_preset_<id>.log.

    • if you use the advanced options form: the log file is named jupyterhub_slurmspawner_advanced_<id>.log. You can even change the name of the output log file (Log File Name).

    • Log File Name also allows you to change the Path of the output log file to a diffrent directory where you have write permissions.

      Example: /path/to/your/scratch/name_of_your_logfile.

  • Use the search field in documentation, or look into the blog posts here, we have written some tutorials on well-known issues

  • Check your job information (services –> accounts), this is useful when the server does not start due to Priority or AssocMaxJobsLimit

../../../_images/squeue.png

If the above steps do not help, send an e-mail to support@dkrz.de.

Known issues#

The following error message means that there is an issue with your EMail address. You need just to update it and contact support@dkrz.de to get access again.

../../../_images/acces_denied.png
  • Consult the documentation or the JupyterHub-related blog posts, which may contain solutions to the issue you are experiencing.” If the above steps do not help, send an e-mail to support@dkrz.de.