JupyterLab is crashing repeatedly and producing a 'SIGILL' error code when I try to run a large notebook on Chrome

[kburchfiel]

Description

I am attempting to run a large notebook (around 1,000 cells) within Chrome. In the process of running the notebook, its Chrome tab crashed around 5 or 6 times, leading to an 'Aw, Snap!' error page with a SIGILL error code.

Has anyone else experienced a similar issue, and if so, what steps would you recommend that I take to resolve it? I tried clearing the output of all of my cells, but this too produced an error message several times.

My laptop has 64 GB of RAM, so if the issue is related to the notebook's exceeding some predefined memory limit, perhaps I could resolve the issue by increasing that limit?

Reproduce

1. Go to '...'
2. Click on '...'
3. Scroll down to '...'
4. See error '...'

Expected behavior

Context

• Operating System and version:
• Browser and version:
• JupyterLab version:

Troubleshoot Output

Paste the output from running `jupyter troubleshoot` from the command line here.
You may want to sanitize the paths in the output.

Command Line Output

Paste the output from your command line running `jupyter lab` here, use `--debug` if possible.

Browser Output

Paste the output from your browser Javascript console here, if applicable.

[welcome]

Thank you for opening your first issue in this project! Engagement like this is essential for open source projects! 🤗

If you haven't done so already, check out Jupyter's Code of Conduct. Also, please try to follow the issue template as it helps other other community members to contribute more effectively.

You can meet the other Jovyans by joining our Discourse forum. There is also an intro thread there where you can stop by and say Hi! 👋

Welcome to the Jupyter community! 🎉

[krassowski]

@kburchfiel which version of JupyterLab did you use?

Your issue might be a duplicate of #15465 which should have been fixed in the latest release, https://github.com/jupyterlab/jupyterlab/releases/tag/v4.1.5, by #15970.

[krassowski]

If you used the latest version/can reproduce it in the latest version, can you share the notebook which reproduces the issue and exact reproduction steps?

[kburchfiel]

According to conda list and the About JupyterLab window within JupyterLab, I'm running version 4.1.5. However, running conda update jupyterlab within my terminal did bring up a number of other libraries to update, so perhaps some of jupyterlab's dependencies weren't up-to-date? I went ahead and updated those libraries, so I'll see if that resolves the issue.

I also tried running conda update jupyterlab within my base environment, but conda ended up hanging indefinitely. I'm not sure why the update command would work for a virtual environment but not a base one, but perhaps this issue is related to my original question.

I'm not able to share this file in particular, but if I see similar behavior within a file that I can share publicly, I'll make sure to do so.

I did try running JupyterLab in Firefox and didn't see the same behavior, so perhaps the issue is related to Chrome?

Update: JupyterLab has been working great on Firefox for the last few hours. This same notebook would often freeze up in both Chrome and Visual Studio Code, especially after it was open for a little while. Clearing my outputs probably helped, but that step wasn't even possible in Chrome, since the page would crash before the outputs had cleared out.

I'd still like to figure out why the program is having issues in Chrome, but if anyone else is experiencing issues with large notebooks, consider running your notebook in Firefox instead.

[krassowski]

so perhaps some of jupyterlab's dependencies weren't up-to-date

No, the fix was in the jupyterlab package itself, so you should be good.

I'm not able to share this file in particular, but if I see similar behavior within a file that I can share publicly, I'll make sure to do so.

That would be appreciated!

I'd still like to figure out why the program is having issues in Chrome, but if anyone else is experiencing issues with large notebooks, consider running your notebook in Firefox instead

This is indeed something that I noted that the crashes only occurred in Chromium based browsers (I saw it in Chrome and Edge). However, I did less testing in Firefox.

Did you use full windowing mode by any chance? (Settings → Notebook → Windowing mode)

[kburchfiel]

Thanks for the prompt response! I'm testing out full windowing mode within JupyterLab on Chrome, and so far the notebook is working great. I'm not seeing any crashes or SIGILL errors appear, and the performance is fine also.

I'll check to see which windowing mode had been set on Firefox, but I believe it was 'defer', since I hadn't made any changes to the default windowing mode there.

So it does seem that the 'defer' windowing mode causes issues in at least some large notebooks in Chrome (but perhaps not Firefox), but these can be resolved by changing the mode to 'full.'

I can of course also try to reduce the amount of output the notebook is presenting; it's not like I'm reviewing all 1000 cells each time I run it. But it looks like this step may not be necessary as long as I keep the full windowing mode in place.

[JasonWeill]

@kburchfiel Thank you for opening this issue! If you could share with us a sample notebook that causes this problem to recur, that would help a developer work on a fix. Even describing the notebook (numbers of cells, outputs, visual outputs, etc) would help give us perspective.

This also represents an opportunity to improve our documentation to suggest size limits for notebooks out of performance concerns. This is likely to vary widely based on notebook type and computer type.

#15258 concerns enabling "full" windowing notebook mode by default (thanks @RRosio )