[issue34745] asyncio ssl memory leak

Alexander Mohr added the comment: going to close, I've verified that it fixes my original issue, ty!! -- resolution: -> fixed stage: patch review -> resolved status: open -> closed ___ Python tracker

[issue34745] asyncio ssl memory leak

Change by Fantix King : -- pull_requests: +12348 ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue34745] asyncio ssl memory leak

miss-islington added the comment: New changeset 7f7485c0605fe979e39c58b688f2bb6a4f43e65e by Miss Islington (bot) in branch '3.7': bpo-34745: Fix asyncio sslproto memory issues (GH-12386) https://github.com/python/cpython/commit/7f7485c0605fe979e39c58b688f2bb6a4f43e65e -- nosy:

[issue34745] asyncio ssl memory leak

Change by miss-islington : -- pull_requests: +12344 ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue34745] asyncio ssl memory leak

Yury Selivanov added the comment: New changeset f683f464259715d620777d7ed568e701337a703a by Yury Selivanov (Fantix King) in branch 'master': bpo-34745: Fix asyncio sslproto memory issues (GH-12386) https://github.com/python/cpython/commit/f683f464259715d620777d7ed568e701337a703a --

[issue34745] asyncio ssl memory leak

Change by Fantix King : -- keywords: +patch pull_requests: +12343 stage: -> patch review ___ Python tracker ___ ___

[issue34745] asyncio ssl memory leak

Change by Alexander Mohr : -- versions: +Python 3.6 ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue34745] asyncio ssl memory leak

Chase N Peterson added the comment: I ran the code snippet below using uvloop/master in a docker container. As it ran, the container continually leaked memory. I included a graph with the memory usage. Environment: # cat /etc/*-release PRETTY_NAME="Debian GNU/Linux 9 (stretch)" NAME="Debian

[issue34745] asyncio ssl memory leak

Change by Fantix King : -- nosy: +fantix ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue34745] asyncio ssl memory leak

Yury Selivanov added the comment: Would you be able to test uvloop/master branch? Current uvloop 0.11.x uses pretty much the asyncio implementation; the master branch has a completely rewritten SSL layer. If the master branch has the leak it might mean that the root cause is indeed in

[issue34745] asyncio ssl memory leak

Alexander Mohr added the comment: sorry, by "raw" I mean in the context of aiohttp, so just using the normal python ssl context and asyncio sockets. I don't think it's an object not getting GC'd because I didn't see any increase on object counts, nor leaks per tracemalloc. I think it's

[issue34745] asyncio ssl memory leak

Yury Selivanov added the comment: What is "raw asyncio SSL sockets"? We don't use SSL sockets in asyncio, we use SSL Memory BIO. Do you think that some SSL context objects aren't being properly released? BTW, can you see the leak when run under uvloop? --

[issue34745] asyncio ssl memory leak

New submission from Alexander Mohr : I've been trying to track down a leak in aiohttp: https://github.com/aio-libs/aiohttp/issues/3010 it seems like this leak now occurs with raw asyncio SSL sockets. when the gist script is run like so: python3.7 `which mprof` run --interval=1