如何修复asyncio中出现的Unclosed connection错误？

相关代码文件 #

fuzzer.py #

import aiohttp
import asyncio


async def get_page(url=None, extensions=None):
    with open("wordlist.txt") as file:
        directories = file.readlines()
    data = []
    async with aiohttp.ClientSession() as session:
        for directory in directories:
            response = await session.get(url=f"{url}/{directory}")
            data.append([f"{url}/{directory}", response.status, response.content_length])
            for extension in extensions:
                response = await session.get(url=f"{url}/{directory}.{extension}")
                data.append([f"{url}/{directory}.{extension}", response.status, response.content_length])
    return data


async def main():
    await get_page()


if __name__ == "__main__":
    asyncio.run(main())

main.py #

from fuzzer import get_page
@dp.message_handler()
async def fuzzer(message: types.Message, state: FSMContext):
    await state.update_data(ex=message.text.split(','))
    data = await state.get_data()
    fuzz_data = await get_page(url=data['url'], extensions=data['ex'])
    cur_time = datetime.datetime.now().strftime('%d_%m_%Y_%H_%M')
    async with aiofiles.open(f"{data['url'][8:]}_{cur_time}.csv", 'w') as file:
        writer = AsyncWriter(file)
        await writer.writerow(
            [
                'URL',
                'Код статуса',
                'Длина контента'
            ]
        )
        await writer.writerows(
            fuzz_data
        )

    await state.finish()

错误日志 #

2022-07-18 04:11:47,729 - ERROR - asyncio - Unclosed connection
client_connection: Connection<ConnectionKey(host='google.com', port=443, is_ssl=True, ssl=None, proxy=None, proxy_auth=None, proxy_headers_hash=None)>

修复方案 #

该错误的核心原因是：获取aiohttp响应后未正确释放响应对象，导致底层连接无法被回收复用。

修改fuzzer.py中的get_page函数，采用async with语法管理每个请求的响应对象，这样代码块结束时会自动关闭响应、释放连接，同时顺便修复readlines()保留换行符导致的URL格式错误：

async def get_page(url=None, extensions=None):
    with open("wordlist.txt") as file:
        # 去除每行末尾的换行符，避免URL拼接后包含无效字符
        directories = [dir.strip() for dir in file.readlines()]
    data = []
    async with aiohttp.ClientSession() as session:
        for directory in directories:
            # 用async with管理响应，自动释放连接
            async with session.get(url=f"{url}/{directory}") as response:
                data.append([f"{url}/{directory}", response.status, response.content_length])
            for extension in extensions:
                async with session.get(url=f"{url}/{directory}.{extension}") as response:
                    data.append([f"{url}/{directory}.{extension}", response.status, response.content_length])
    return data

如果不想使用async with，也可以在获取响应后手动调用await response.release()来释放连接，但async with是更简洁、安全的写法。

内容的提问来源于stack exchange，提问作者spot