Giters

observablehq / feedback

Customer submitted bugs and feature requests

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Downloading large attachments from Google stalls

boxfoot opened this issue · comments

commented

Describe the bug
A clear and concise description of what the bug is.

Observable frequently (but not always) stalls out when loading large cloud attachments from Google. This regularly happens to me with attachments larger than ~50MB. Sometimes they load normally and everything is fine. Other times, the download seems to stall and Observable remains in a loading state indefinitely.

I've noticed that this seems to happen day by day -- on some days, everything will load as expected (no matter how many times I reload a notebook) and other days, the stalling behavior will happen consistently (no matter how many times I load the behavior).

The stalling can be watched by opening the chrome developer console, going to the network tab, and filtering requests for those named file?token= .... Initially, this load as expected, and then stop suddenly when the file has not been fully downloaded.

To Reproduce
Steps to reproduce the behavior:

  1. Upload a large (~100MB) csv file to google drive. Make sure it's a CSV file, not a google sheet.
  2. Create a new notebook and pull in that google drive file.
  3. Add a cell to the notebook, something like FileAttachment("google://AllRecords.csv").text()
  4. Cell never loads (except when it does)

Expected behavior
Attachment should load and be available for use in sheet.

Screenshots

Showing loading status after several minutes of sitting:
image

Network panel showing two stalled attempts to load the file:
image

Details from one of the requests:

image

Desktop (please complete the following information):

  • OS: iOS
  • Browser arc and chrome (happens in both)
  • Version 115

Additional context
I wonder if this is related to #576 . And it would be much less of a pain if #580 or similar were available, because we'd be less dependent on downloading files repeatedly.

commented

UPDATE: One more observation: I sometimes work through a VPN and sometimes without, and I've noticed this issue is more likely to happen on days when I am working through the VPN.

commented

I have been testing this with a 5,000,000 row csv file on Google Drive (111MB). It does seem to load every time for me, but sometimes it is slower than other times. When we looked at our logs, we did notice your incomplete cloud file requests with an error "Client Request Interrupted”. I assume that is when you either refreshed or closed your browser? Wondering whether it would show up if you just wait long enough? For my test, I had to sometimes wait for a whole minute for the file to download...

commented

@CobusT Thanks for looking into it. I am not aware of actions I took on my side to interrupt the request -- and I sometimes would leave it sitting for 10-30 minutes to see if it would resume. But like I said in the post it's really uneven -- some days no issues and other days a lot. Today is a "everything is working" day for me...

commented

Closing this issue for now, but feel free to re-open if discover any patters that you think could be causing this intermittent problem.