Django/Daphne/Nginx Websockets 未连接

Question

我正在使用Django和Daphne和Nginx来运行我的Websocket。几个月来，一切都很好，但是几天前，我的Websockets不会连接说“ Websocket连接到'wss：// mydomain：8001/clips/clips/五十9/'失败：”。

我检查了服务器日志，这是在输出中，

2022-04-07 16:57:22,970 ERROR    [Failure instance: Traceback: <class 'AttributeError'>: 'NoneType' object has no attribute 'replace'
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/autobahn/websocket/protocol.py:2839:processHandshake
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/txaio/tx.py:366:as_future
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/twisted/internet/defer.py:151:maybeDeferred
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/daphne/ws_protocol.py:72:onConnect
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: --- <exception caught here> ---
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/twisted/internet/defer.py:151:maybeDeferred
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/daphne/server.py:200:create_application
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/channels/routing.py:54:__call__
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/channels/security/websocket.py:35:__call__
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/channels/security/websocket.py:53:valid_origin
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/channels/security/websocket.py:72:validate_origin
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/channels/security/websocket.py:73:<genexpr>
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /home/django/Odyssey/venv/lib/python3.8/site-packages/channels/security/websocket.py:97:match_allowed_origin
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /usr/lib/python3.8/urllib/parse.py:376:urlparse
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: /usr/lib/python3.8/urllib/parse.py:430:urlsplit
Apr 07 16:57:22 ubuntu-s-1vcpu-1gb-nyc1-01 python[1276]: ]

我不明白为什么突然发生这种情况，这可能是依赖性被更新还是弃用了？

Answer 1

这是因为 urllib3 中的修复。允许的霍斯特罗素参与者是经纪的。我解决了此编辑的 asgi.py 文件。

我从此更改了“应用程序”定义：

import os

from .wsgi import *
from channels.auth import AuthMiddlewareStack
from channels.security.websocket import AllowedHostsOriginValidator
from channels.routing import ProtocolTypeRouter, URLRouter
from django.core.asgi import get_asgi_application

os.environ.setdefault(
    'DJANGO_SETTINGS_MODULE',
    'manager.settings.production'
)

from .routing import websocket_urlpatterns
application = ProtocolTypeRouter({
    "websocket": AllowedHostsOriginValidator(
        AuthMiddlewareStack(
            URLRouter(
                websocket_urlpatterns
            )
        )
    ) 
})

到此处：

import os

from .wsgi import *
from channels.auth import AuthMiddlewareStack
from channels.security.websocket import AllowedHostsOriginValidator
from channels.routing import ProtocolTypeRouter, URLRouter
from django.core.asgi import get_asgi_application

os.environ.setdefault(
    'DJANGO_SETTINGS_MODULE',
    'manager.settings.production'
)

from .routing import websocket_urlpatterns
application = ProtocolTypeRouter({
    "websocket": AuthMiddlewareStack(
        URLRouter(
            websocket_urlpatterns
        ) 
    )
})

更多信息： https：// github .com/django/channels/essess/1716

Answer 2

我通过编辑redis.service文件解决了这个问题

nano /etc/systemd/system/redis.service

execstop =/bin/kill -s term $ mainpid

execstartpost =/bin/sh -c“ echo $ mainpid＆gt;/var/run/redis/redis.pid”

然后键入

sudo systemctl守护程序 - 雷亚德

sudo systemctl启用redis服务器

sudo systemctl restart redis.service