code-server: Jupyter Notebook cells don’t execute properly

OS/Web Information

  • Web Browser: Brave
  • Local OS: macOS
  • Remote OS: docker on macoS
  • Remote Architecture: amd64
  • code-server --version: 3.9.3

Steps to Reproduce

  1. Click a .ipynb file

Expected

Jupyter Notebook should open, connect to a Kernel and cells should be ready to run

Actual

Even after a series of installations of plugins (that seem to work fine) the Notebook can connect to a kernel but no output is being received (all i see is “Loading…”, then after some time I get the error shown in the screenshot “ChunkLoadError”)

Screenshot

Screenshot 2021-04-23 at 20 50 21

Notes

This issue can be reproduced in VS Code: No idea.

About this issue

  • Original URL
  • State: closed
  • Created 3 years ago
  • Comments: 27 (10 by maintainers)

Most upvoted comments

For completeness: after resetting Brave this issue is gone even when I use the Marketplace Extensions Jupyter v2021.3.0 and Python v2021.4.765268190 which are being installed as defaults (code server 3.9.3)!

I was trying to provide the console logs when the bug just went away. I don’t know why but it’s not an issue any more. I’m closing this, sorry for the fuzz.

Very interesting! I’m glad we got it sorted! 😄

Thanks, @code-asher your help solved my issue!

After deleting all user data related to Brave and re-installing it cell execution works again like it should!

This is so weird, I was trying

docker build -t jupyter . docker run -v $(pwd):/app -p 8080:8080 -it -d jupyter docker exec -it <dontainer_id> cat /root/.config/code-server/config.yaml (to see the generated password)

Then follow 4, 5, 6, 7, 8, 9, 10, 11 and 12

And all I got is “Loading…” and then the ChunkLoadError. That was quite unexpected because we were using the same Dockerfile. My Docker host is macOS Big Sur 11.3; But my Broser is Brave - so i tried Safari and: The cell executed!

I gave your Dockerfile a run and everything worked! My exact steps:

  1. Save Dockerfile as test.Dockerfile
  2. docker build -t code-server . -f test.Dockerfile
  3. docker run -e PASSWORD=test -p 8080:8080 code-server
  4. Browse to localhost:8080 and log in
  5. Open terminal
  6. Run wget https://open-vsx.org/api/ms-python/python/2020.10.332292344/file/ms-python.python-2020.10.332292344.vsix
  7. Extensions > Install from VSIX > Select downloaded VSIX
  8. Refresh the page
  9. Open command palette
  10. Select Create New Blank Jupyter Notebook
  11. Add print("hello") to the cell
  12. Click run

I gave it a try and I’m seeing the same thing (or something similar). It says Not connected. If I click it to connect it fails with:

Error: Jupyter cannot be started. Error attempting to locate jupyter: Data Science libraries jupyter and notebook are not installed in interpreter Python 2.7.17 64-bit

If I switch to Python 3 using the button in the status bar I get this instead:

Error: Jupyter cannot be started. Error attempting to locate jupyter:

Jupyter is in my path so I’m not sure why.

So then I opened the command pallete and switched the interpreter there instead and for some reason that worked. (Credit here: https://stackoverflow.com/questions/60330837/jupyter-server-not-started-no-kernel-in-vs-code)

I don’t know if any of that applies to your situation but maybe it’ll be helpful.

I tried to replicate but I ran into issues with the extension activating at all; I’m getting the error mentioned here: https://github.com/cdr/code-server/issues/2929

I’d be interested in seeing if downloading and installing the extension from OpenVSX instead of our marketplace works for you: https://open-vsx.org/extension/ms-python/python