[BUG] - playwrights runs in multitudes

My code is essentially is a loop over

tt = api.getTikTokById(video['id'],proxy=proxy,custom_did=custom_did)
bytes = api.get_Video_By_TikTok(tt['itemInfo']['itemStruct'],proxy=proxy,custom_did=custom_did)

api get created in the begin of the program.

Due to some reason, playwrights did not stop and runs until the program finishes. They merrily eats all CPU and memory, at number of 20-30 they causes my machine to hang.

Actually, after some processing, it yields

6816498522741918981 info...loading...6816498522741918981 is video/mp4...done.
6813478929186245893 info...loading...Error at 90 Protocol error (Runtime.enable): The page has been closed. loading 6813478929186245893 error.
6830234382377258245 info...Error at 90 Target page, context or browser has been closed loading 6830234382377258245 error.

And after stop I still see

tarkhil     9917  0.0  0.0   6892  1704 ?        SNs  14:03   0:00 /bin/bash /home/tarkhil/.cache/ms-playwright/webkit-1383/pw_run.sh --inspector-pipe --headless --no-startup-window --proxy=socks5://148.xxx.xxx.247:1080
tarkhil     9922  0.5  0.0 104291544 3176 ?      SNl  14:03   0:12 /home/tarkhil/.cache/ms-playwright/webkit-1383/minibrowser-wpe/bin/MiniBrowser --inspector-pipe --headless --no-startup-window --proxy=socks5://148.xxx.xxx.247:1080

Should I reinint API on each run and kill all subprocesses after?..

[zoopzorp]

This happens because a new BrowserContext is created upon each request and never closed. I am not sure if a new context is needed each time or not, but to fix the current issue keeping functionality the same:

• Return the context from create_page instead of a page
• Change the two calls to create_page to instead create a page afterwards with the returned context
• Call context.close() after the operation finishes instead of page.close() (You will have to reorder the operations in sign_url since the current page.close() is after the return statement)

[pibvbp]

great - thanks!!