[issue31096] asyncio.stream.FlowControlMixin._drain_helper may lead to a blocking behavior

[Yury Selivanov]

Yury Selivanov  added the comment:

Documenting this would be a great first step.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue31096] asyncio.stream.FlowControlMixin._drain_helper may lead to a blocking behavior

[Dmitry Malinovsky]

Dmitry Malinovsky  added the comment:

I've changed the implementation significantly since August 2017, futures are 
not involved anymore so please ignore that part. However, such drain behavior 
is still present - but I don't think anymore that yielding to the event loop is 
an easy fix.

I've tried to do so in my lib, and it showed significant slowdowns (around 4-5k 
publishes per second). It's not acceptable. I also found this message from 
Guido https://github.com/python/asyncio/issues/263#issuecomment-142702725.

What really helped is a counter that tracks send calls without waiting for 
replies, and a user-provided limit; when the counter reaches the limit, an 
explicit yield (via await asyncio.sleep(0)) is performed. This helped to 
achieve around 15-16k publishes per second (3-4 times faster. Here's the code:
https://github.com/malinoff/amqproto/blob/6568204b539ecf820af2da11bddcca9ce7323ac5/amqproto/adapters/asyncio_adapter.py#L53-L71

Now I'm thinking that such behavior should only be documented - so library 
authors can deal with it before they face this in production. But if you have 
other thoughts, I'd be glad to hear.

--

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue31096] asyncio.stream.FlowControlMixin._drain_helper may lead to a blocking behavior

[Yury Selivanov]

Yury Selivanov  added the comment:

Andrew, what are your thoughts on this one?

--
nosy: +asvetlov
versions: +Python 3.8 -Python 3.5, Python 3.6

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue31096] asyncio.stream.FlowControlMixin._drain_helper may lead to a blocking behavior

[Greg Lindahl]

Change by Greg Lindahl :


--
nosy: +wumpus

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue31096] asyncio.stream.FlowControlMixin._drain_helper may lead to a blocking behavior

[Dmitry Malinovsky]

New submission from Dmitry Malinovsky:

When there is a huge amount of `writer.write` calls followed by `await 
writer.drain()` on a non-paused channel, and there are no other coroutine 
switches, `await writer.drain()` immediately returns without a switch. This is 
because `asyncio.stream.FlowControlMixin._drain_helper` do not `yield` or 
`yield from` on a non-paused stream.

Use-case: AMQP basic.publish method, for which the broker (rabbitmq) do not 
send any replies back. Trying to publish 4k messages results in the following 
warnings (PYTHONASYNCIODEBUG env variable is set):
`Executing () created at 
/Users/malinoff/Projects/ideas/amqproto/amqproto/channel.py:85> took 2.371 
seconds`
2.371 seconds is the time spent on 4k `basic_publish` calls.

You can find the test itself on github: 
https://github.com/malinoff/amqproto/blob/master/tests/stress/test_4k_msgs.py#L11-L12

An easy fix would be to replace return 
(https://github.com/python/cpython/blob/master/Lib/asyncio/streams.py#L206) 
with yield (and but the code below under the else clause; I'm willing to 
prepare a pull request), but maybe I'm missing something and such behavior is 
intentional?

--
components: asyncio
messages: 299610
nosy: Dmitry Malinovsky, yselivanov
priority: normal
severity: normal
status: open
title: asyncio.stream.FlowControlMixin._drain_helper may lead to a blocking 
behavior
versions: Python 3.5, Python 3.6, Python 3.7

___
Python tracker 

___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com