Python 3 asyncio how to properly close client connection
I found a memory leak in a tornado web app and I'm not sure how to fix it. After some memory profiling, I found that my memcached client loses the dict when closing client connections. I am opening / closing memcached clients (once a minute to be specific) as part of the w / AWS ElastiCache auto-discovery mechanism .
Below is a minimal projector using pympler to demonstrate the leak:
from pympler import muppy, summary
import asyncio
import aiomcache
loop = asyncio.get_event_loop()
async def hello_aiomcache():
mc = aiomcache.Client("127.0.0.1", 11211, loop=loop)
await mc.set(b"some_key", b"Some value")
value = await mc.get(b"some_key")
print(value)
values = await mc.multi_get(b"some_key", b"other_key")
print(values)
await mc.delete(b"another_key")
mc.close()
# establish a baseline (watch the <class 'dict line)
summary.print_(summary.summarize(muppy.get_objects()))
for i in range(50):
loop.run_until_complete(hello_aiomcache())
# <class 'dict grows
summary.print_(summary.summarize(muppy.get_objects()))
ds = [ao for ao in muppy.get_objects() if isinstance(ao, dict)]
# leaked dict looks like {'_loop': <_UnixSelectorEventLoop running=False closed=False debug=False>, '_paused': False, '_drain_waiter': None, '_connection_lost': False, '_stream_reader': <StreamReader t=<_SelectorSocketTransport fd=34 read=polling write=<idle, bufsize=0>>>, '_stream_writer': None, '_client_connected_cb': None, '_over_ssl': False}
ds[2364]
It looks like this dict will hang forever until it is called loop.close()
. I am confused by this. I think I don't want to close the loop that I borrowed from the tornado through tornado.ioloop.IOLoop.IOLoop.current().asyncio_loop
. Is there any other way to properly close / clear these connections without closing the loop?
source to share
The problem is not caused by await
ing mc.close()
.
I was a little surprised to learn that a coroutine won't work without some explicit scheduling. I naively thought that at some point in the future it would be called. However, the accompanying documents clearly state:
A coroutine call does not start from its code - a coroutine object returned by the call does nothing until you schedule it to execute. There are two main ways to run it: call
await coroutine
oryield from coroutine
from another coroutine (assuming another coroutine is already running!), Or schedule it to run using the security_future () or AbstractEventLoop.create_task () method.
source to share