VNC Desktop

With this feature you can start a VNC Desktop on Levante nodes using the jupyterlab interface.

Note

This feature is experimental and currently under active testing, we do not guarantee a full time availability. We still need to fix some issues. Please test it and report any problem to support@dkrz.de.

Overview

../../../_images/vnc.png

How to start the VNC server

Before starting jupyterlab

This feature is available only for the advanced spawner. From the field User Interface list you select VNC Desktop.

../../../_images/vnc_launcher.png

A VNC desktop will start in a new tab. This server will be stopped when you stop your Jupyterhub session.

After starting jupyterlab

Click on the icon VNC Desktop.

../../../_images/vnc-launcher.png

How to (correctly) stop the VNC server

  • It is important that you stop your jupyterhub session when you finish your work, the VNC Desktop will be stoped automatically.

  • Closing the window/tab does neither stop the VNC Desktop nor the jupyterhub session.

  • If you started the VNC Desktop using the advanced spawner. you need to modify the url (replace /desktop/ by /lab/tree/) to go back to jupyterlab and from there you can stop the VNC Desktop.

How to start software

After starting the VNC Desktop, you can now launch your favorite software by using the terminal.

Example

Note

More modules/packages/configurations can be required to start and work with a specific software.

  • ParaView

Currently, use spack.

$ spack load paraview
$ paraview
  • Matlab

$ module load matlab
$ matlab

Best practices

  • In order to avoid display conflicts, it is advisable to check for running vncservers before starting the VNC Desktop.

vncserver -list

TurboVNC sessions:

X DISPLAY #     PROCESS ID      NOVNC PROCESS ID
  • If you see an error message like this:

WARNING: l*.atos.local:1 is taken because of /tmp/.X11-unix/X1
Remove this file if there is no X server l*.atos.local:1

so remove it first.

Troubleshooting

  • Look into the log files in $HOME/.vnc/

  • Notebook Log file jupyterhub_slurmspawner…..log

  • If you see and error like: subprocess killed, it often means a lack of memory. For e.g. Matlab requires at least 4/5 GB to start