Forums

"can't start new thread" in ipynb using futures.ThreadPoolExecutor

I have a script that I've been using for some time now which constructs a pandas dataframe from the data stored in several different binary files, and does so by using the futures.ThreadPoolExecutor to save time. I have it set up to use a maximum of 10 workers. For months, it had been working as intended, but now I'm getting RuntimeError: can't start new thread, but when I look at my active thread count with threading.active_count() I only see 6 threads. (Been using python 3.5 the whole time). Did something change?

 

[ As additional info, I have a different command line script which also uses futures.ThreadPoolExecutor and I've not encountered this problem with that script, also using python 3.5 ]

I wonder if you have a lot of notebooks open, and the total thread count has increased past our set limit.

If you go to the consoles page and hit the fetch processes, you should be able to kill a bunch of notebook processes. See if that helps?

Aha, that looks like it did the trick. Thanks for the help.