-
Notifications
You must be signed in to change notification settings - Fork 765
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Frequent crashes when using jupyter kernels #6289
Comments
What's your version of the Python extension? This might be a dupe of this issue: which required an update to the Python extension. |
This is the commit in the Python extension that fixed the problem: |
Will check this, thanks. |
OK, we updated to the following:
Can you confirm these have the fix that you mentioned above? It was unclear to me if they did. And we're still getting an error that looks similar to the originally reported one. The small piece of additional information is that it seems to happen "after a while" and is triggered when the user has 2 interactive windows open (each connected to a different kernel), and the switches between them.
With regards to trace logging, we have |
looks like we have some issue with interactive window. I have another report on it - #6312 - that probably due to crash as well. |
Let me know if there is anything else I could provide to help. |
Also got a consistent repo. Working with a Jupyter Interactive Window on a Dev Container. Version: 1.93.1 (user setup) |
Interactive window crashes are likely this issue here: #6357. That should be fixed in our latest prerelease build. |
@dlaflamme, are you still seeing this issue on our latest prerelease build? |
No, haven't experienced this in the latest prerelease build. Is this change in a non-prerelease build yet? Thanks for following up. |
Not yet. We should be releasing a new stable build this coming week. Closing as a duplicate of #6357. |
Environment data
Additionally, in this case VS Code is running on a Windows client and the Remote-SSH extension is being used to access python code on the Linux server.
Code Snippet
There is not a specific code snippet that will trigger this. Instead, see the repro steps. he issue seems to be triggered by using the setup described in the repro steps for some period of time.
Repro Steps
Run a Jupyter notebook outside of VS Code. Do some development in it. In VS Code, start a Jupyter interactive session. "Select Another Kernel", then "Existing Jupyter Server", then select the the kernel running in the Jupyter notebook running outside of VS Code. Test it. (for example, set a variable value on one side and make sure you can see it on the other).
Do regular development in the notebook and run things in the interactive session to examine variables set in the Jupyter kernel.
Eventually pylance will crash. The crashing seems to have increased in frequency in more recent versions of Pylance. I have downgraded to 2024.6.1 and the frequency of occurrence had decreased (but has not been eliminated).
Logs
The above shows representative output. There are more lines like above but they are all of the same flavor as you see above.
Then, the last line in that sequence:
Then several blank lines that just look like the following (a log header with no content)
and finally:
Additionally, I have tried setting
"python.analysis.logLevel": "Trace"
(on VS Code GUI side that is running on windows--but I don't see trace level logs in the output on the Linux host, so not sure if I did that right.The text was updated successfully, but these errors were encountered: