[Twisted-Python] twisted thread pool

[Al Niessner]

Platform:
CPU: Intel (2x with 8 core each then hyper-threading gets us to 32 virtual)
Ubuntu 14.04
Python 3.4 (yes, I know)
Twisted 13.2.0

I use the vanilla twisted.internet.reactor.run() with the
factory twisted.web.server.Site(root). It all runs just fine (seemingly)
but,when I was snooping around on my machine, I had a ton of threads lying
about. After cleaning up, I restarted my twisted application and found that
I have 151 python3 threads lying about. I doubled checked and I have just
twisted application running.

I browsed the documentation and the twisted thread pool claims to default
between 5 and 10. However, I do abuse the deferToThread() cannot find any
simple documentation on how that might feed back to the 151 threads. I
thought maybe I need to exit the deferred returned from deferToThread() in
some special way to clean it all up. However, I can find no documentation
for it.

I have been googling about with many random search terms trying to puzzle
the 151 threads from the documentation that tells me I should expect 5 to
10.

Can anyone please help understand why I have 151 threads and where they are
coming from? Luckily they are all idle so the machine runs just fine.

The best story I can come up with is that there is one master thread on
core-0 and then I get 10 threads in the pool for each of the other physical
cores. However, the story is just numerology.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://twistedmatrix.com/pipermail/twisted-python/attachments/20150302/28a3c87d/attachment.html>