Catalyst connection freezes in Jupyter

I am trying to connect from Jupyter to a catalyst in situ simulation, but Jupyter freezes in the moment, when the in situ side connects. ( The aim of this experiment is to show some statistics, no graphics, in Jupyter from the insitu side.)

What I did:

  • Installed the iparaview kernel in Jupyter, works without problems
  • Run a modified version of Examples/PythonClient/catalystLiveVisualizer.py in Jupyter
  • Start ./CxxFullExample feslicescript.py as suggested in the above example
  • at this moment, the _monitorServerNotifications() loop stops

I attach my modified version of Examples/PythonClient/catalystLiveVisualizer.py with the following changes:

  • all Rendering related calls removed
  • _monitorServerNotifications(), _updateEventCb(), _connectionCreatedEventCb() print a message

When I start the attached script in pvpython, it works as expected, shows “waiting…” repeatedly, then the outputs from the connectionCreated and updateEvent callbacks after I start the insitu side.

When I start the script in Jupyter, it shows

Accepting connection(s): myhost:22222
waiting...
waiting...
waiting...
...

… and the output stops when I start the insitu side.

Questions:

Do you have any idea, what interrupts _monitorServerNotifications()? I also wrote the messages to a file to be sure that not the Jupyter output capturing is the problem.

How can I start debugging this problem?

I know that the problem may be the Jupyter side. But I assume that many of the Paraview community - at least the iparaview-kernel developers - have more insight into Jupyter than most Jupyter developers know about Paraview.

Thanks, Gregor

catalystLiveVisualizer-debug.py (1.8 KB)