Lots of jedi subprocesses created until memory runs out

[dorellang]

It seems to be a regression of #3586 and #3514 . This started to happen recently (I can't tell exactly when, but a couple of weeks back I didn't experience this issue).

Environment data

• VS Code version: 1.37.1
• Extension version (available under the Extensions sidebar): 2019.8.30787
• OS and version: MacOS High Sierra Version 10.13.6
• Python version (& distribution if applicable, e.g. Anaconda): Python 3.7.3
• Type of virtual environment used (N/A | venv | virtualenv | conda | ...): virtualenv
• Relevant/affected Python packages and their versions: XXX
• Jedi or Language Server? (i.e. what is "python.jediEnabled" set to; more info How to update the language server to the latest stable version #3977): Jedi

Actual behaviour

After working for a while in a Django project, python completion.py spawns lots of subprocesses that are never killed. Eventually the system runs out of memory and fork fails. You can close the visual studio folder in order to free memory, but it takes a while for the python processes to be killed and the computer gets unresponsive.

Steps to reproduce:

1. Open a python project
2. Browse some files, and on each of them call autocomplete and save several times. After a while you should get a lot processes that won't go away.

[karrtikr]

Browse some files, and on each of them call autocomplete and save several times. After a while you should get a lot processes that won't go away.

Can you please provide the file content where you call autocomplete, not aware of it. Thanks!

[wj-c]

I have the same issue. Lots of processes are spawned after using the extension for a while.

[karrtikr]

@wj-c Processes maybe spawned, but are those being killed when you stop using the extension?

[wj-c]

No, they are not killed after I completely close vscode and python extension.

…

-------- 原始信息 -------- 发件人： Kartik Raj <[email protected]> 日期: 2019/9/14 03:08 (GMT+08:00) 收件人： microsoft/vscode-python <[email protected]> 抄送： wj-c <[email protected]>, Mention <[email protected]> 主题： Re: [microsoft/vscode-python] Lots of jedi subprocesses created until memory runs out (#7144) @wj-c<https://github.com/wj-c> Processes maybe spawned, but are those being killed when you stop using the extension? ― You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#7144?email_source=notifications&email_token=AIO72JHIM5R75RUD2UOXEMLQJPQMNA5CNFSM4ISDCZQ2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD6V5XOQ#issuecomment-531356602>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AIO72JBHX4ITGCJ4JODCRULQJPQMNANCNFSM4ISDCZQQ>.

[deitry]

Hi.
There is also similar problem for pylint: #4236
I found recently that processes do not die automatically when spawned for files with errors. Maybe there is something common in these issues?

[karrtikr]

@wj-c Apologies, not sure what might be happening there. Are child process being created in the python code? I am not able to reproduce this just with any code. Please try to provide a reproducible case and create a new issue as it might be a different one.

@deitry Thanks for pointing to that. Unfortunately without a reproducible case, there's not much we can do here.

[karrtikr]

Closing the original issue for now as no further response was provided by @dorellang . I'll happy to reopen and address this if further clarification is provided.

[deitry]

@karrtikr
See #4236 (comment)
At least for me with pylint - described way reproduces issue 10 times of 10.

[DonJayamanne]

@deitry
Thanks, however that's a separate issue even though they may seem related.
We'll monitor/track that issue separately. If indeed they end up being related (generic issue), then we'll end up fixing this as well.