Skip to content
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

Long Pauses Rendering Notebooks On Initial Load with Code-Server #652

Open
jlewi opened this issue Aug 9, 2024 · 6 comments
Open

Long Pauses Rendering Notebooks On Initial Load with Code-Server #652

jlewi opened this issue Aug 9, 2024 · 6 comments

Comments

@jlewi
Copy link
Contributor

jlewi commented Aug 9, 2024

I started noticing that when I would try to open a notebook it would take a really long time to render or might not render at all.
I'll have a couple notebooks open and then try to open a new notebook and it will take a long time (10s of seconds -minutes) to render.

I'm running runme in code-server and then accessing it in the browser. So it could be particular to that setup. I haven't tried to see if this problem occurs in other settings.

It seemed to be working fine for most of the day and then this problem started occurring.

I can open the file in the markdown editor just fine. So my suspicion is it has something to do with serialization.
The next time I observe this I'll try clearing the app data in the browser and seeing what happens.

I've tried reloading the window and that doesn't seem to help.

@sourishkrout
Copy link
Member

Would you mind adding a video clip @jlewi? I just tried it myself, and it opened instantly.

The very first open (and every 30d) might take a few seconds longer (entropy for pk) because Runme is rotating its CA. Sounds like that is not it in your case.

@jlewi
Copy link
Contributor Author

jlewi commented Aug 9, 2024

Will do. FWIW, I recognize this is an unhelpful bug report and don't know how to reproduce it. I think its reasonable to close it as unactionable. I mainly wanted to start a record where I could start to log information.

@jlewi
Copy link
Contributor Author

jlewi commented Aug 10, 2024

Here's a video; I had to record with GVC so its not as nice as with screen recorder.

Here's what the video is showing

  • I just opened a brand new notebook build_tensorrt.md and its "spinning" trying to load it
  • I try to click on the buttons to add a code cell or markdown cell and nothing happens
  • I click on the M window and instead of seeing the markdown for the new notebook I get a different file
  • I then right click and say open in text editor for build_Tensorrt.md
  • Interestingly, the raw markdown shows a bunch of code cells
    • Presumably these are the result of click on the add code cell button; but they aren't rendering.
runmebug-notebook-doesnt-render.mp4

@jlewi
Copy link
Contributor Author

jlewi commented Aug 11, 2024

Here's another example.

  • When the video starts I have an open file build_tensorrtllm.md which isn't rendering
  • I click on M and it doesn't show up
  • I then close it and reopen it and it doesn't render
  • When I look at the RunMe logs we see a bug from the GrpcProgramRunner about not being to write to a closed program session
  • I reset the session and try again and it seems to render.
runmebug-notebook-doesnt-render-existing-file.mp4

@sourishkrout
Copy link
Member

sourishkrout commented Aug 12, 2024

Are there any suspicious-looking logs in Output->Runme (bottom panel)? My guess is that they're related to task discovery, where we traverse the repo's tree for markdown files.

Perhaps related to this #646.

@sourishkrout
Copy link
Member

I just remember that I've seen issue like this on Remote Workspace when I have a local runme binary set in my extension settings, such as:

"runme.server.binaryPath": "/Users/sourishkrout/Projects/stateful/oss/runme/runme"

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants