Socket.io 远程但不是本地时回退到 nginx 代理后面的轮询(websocket 提供 400)

Question

我正在使用nginx 1.4.6作为Django应用程序和使用 socket.io 1.2.1以及socketio-jwt 2.3.5的相关NodeJS应用程序的代理

在本地，我使用 Vagrant 实例 (ubuntu 14.04.1( 进行编码，而远程我使用的是专用的 AWS EC2 实例 - 两者都使用相同的 Vagrant file 和 setup shell 脚本进行配置，因此它们非常接近相同。

在本地运行时，我的客户端连接到nginx服务器，该服务器升级连接并将其传递给Node下的 socket.io 服务器实现。一切都很好。

远程地，我记录了与transport=polling的第一个连接，这也指示客户端升级到 websocket 传输，然后是第二个请求transport=websockets，然后是第三个请求，再次是transport=polling。

虽然应用程序仍然可以工作，但出于多种原因

第二个请求的回答是400 Bad Request。使用 DEBUG=* node index.js 运行我的 Node 应用会给出以下输出：

engine intercepting request for path "/socket.io/" +24s
engine handling "GET" http request "/socket.io/?token=eyJhbGciOiJIUzUxMiIsInR5cCI6IkpXVCJ9.eyJ1c2VybmFtZSI6Im1hcmNlbCIsIm9yaWdfaWF0IjoxNDE4MDU5ODg1LCJ1c2VyX2lkIjozLCJlbWFpbCI6IiIsImV4cCI6MTQxODE0NjI4NX0.2PE0TNRol9G4hby4OzQ-af2e0yjfgFAb-gQJF5tKWRwxWnFLv1NGp3Yo87UaqNaQceMW6KqzMIx2gLcRFnk09A&EIO=3&transport=polling&t=1418062322853-0" +0ms
engine handshaking client "6YDFOjBfvZ9UyutVAAAB" +1ms
engine:socket sending packet "open" ({"sid":"6YDFOjBfvZ9UyutVAAAB","upgrades":["websocket"],"pingInterval":25000,"pingTimeout":60000}) +0ms
engine:polling setting request +0ms
engine:socket flushing buffer to transport +0ms
engine:polling writing "  �0{"sid":"6YDFOjBfvZ9UyutVAAAB","upgrades":["websocket"],"pingInterval":25000,"pingTimeout":60000}" +1ms
engine:socket executing batch send callback +1ms
socket.io:server incoming connection with id 6YDFOjBfvZ9UyutVAAAB +1.7m
socket.io:client connecting to namespace / +1.7m
socket.io:namespace adding socket to nsp / +1.7m
socket.io:socket socket connected - writing packet +1.7m
socket.io:socket joining room 6YDFOjBfvZ9UyutVAAAB +0ms
socket.io:client writing packet {"type":0,"nsp":"/"} +79ms
socket.io-parser encoding packet {"type":0,"nsp":"/"} +1.7m
socket.io-parser encoded {"type":0,"nsp":"/"} as 0 +0ms
engine:socket sending packet "message" (0) +79ms
socket id: "6YDFOjBfvZ9UyutVAAAB" connected to user: "marcel"
socket.io:socket joined room 6YDFOjBfvZ9UyutVAAAB +1ms
engine intercepting request for path "/socket.io/" +121ms
engine handling "GET" http request "/socket.io/?token=eyJhbGciOiJIUzUxMiIsInR5cCI6IkpXVCJ9.eyJ1c2VybmFtZSI6Im1hcmNlbCIsIm9yaWdfaWF0IjoxNDE4MDU5ODg1LCJ1c2VyX2lkIjozLCJlbWFpbCI6IiIsImV4cCI6MTQxODE0NjI4NX0.2PE0TNRol9G4hby4OzQ-af2e0yjfgFAb-gQJF5tKWRwxWnFLv1NGp3Yo87UaqNaQceMW6KqzMIx2gLcRFnk09A&EIO=3&transport=polling&t=1418062323048-1&sid=6YDFOjBfvZ9UyutVAAAB" +0ms
engine setting new request for existing client +1ms
engine:polling setting request +0ms
engine:socket flushing buffer to transport +1ms
engine:polling writing "�40" +38ms
engine:socket executing batch send callback +1ms
engine intercepting request for path "/socket.io/" +0ms
engine handling "GET" http request "/socket.io/?token=eyJhbGciOiJIUzUxMiIsInR5cCI6IkpXVCJ9.eyJ1c2VybmFtZSI6Im1hcmNlbCIsIm9yaWdfaWF0IjoxNDE4MDU5ODg1LCJ1c2VyX2lkIjozLCJlbWFpbCI6IiIsImV4cCI6MTQxODE0NjI4NX0.2PE0TNRol9G4hby4OzQ-af2e0yjfgFAb-gQJF5tKWRwxWnFLv1NGp3Yo87UaqNaQceMW6KqzMIx2gLcRFnk09A&EIO=3&transport=websocket&sid=6YDFOjBfvZ9UyutVAAAB" +1ms
engine bad request: unexpected transport without upgrade +0ms
engine intercepting request for path "/socket.io/" +120ms
engine handling "GET" http request "/socket.io/?token=eyJhbGciOiJIUzUxMiIsInR5cCI6IkpXVCJ9.eyJ1c2VybmFtZSI6Im1hcmNlbCIsIm9yaWdfaWF0IjoxNDE4MDU5ODg1LCJ1c2VyX2lkIjozLCJlbWFpbCI6IiIsImV4cCI6MTQxODE0NjI4NX0.2PE0TNRol9G4hby4OzQ-af2e0yjfgFAb-gQJF5tKWRwxWnFLv1NGp3Yo87UaqNaQceMW6KqzMIx2gLcRFnk09A&EIO=3&transport=polling&t=1418062323180-2&sid=6YDFOjBfvZ9UyutVAAAB" +1ms
engine setting new request for existing client +0ms
engine:polling setting request +0ms

这个问题似乎与线路engine bad request: unexpected transport without upgrade +0ms有关，但我不明白。nginx配置肯定提到了升级，它可以在我的流浪机器上运行。

nginx 配置的相关部分如下：

server {
    ...
    location / {
        ...
    }
    location /socket.io/ {
        proxy_pass http://127.0.0.1:4000/socket.io/;
        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "upgrade";
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-NginX-Proxy true;
        proxy_redirect off;
    }
}

我能想到的本地和远程设置之间的唯一区别是，在本地，客户端在两个不同的端口上连接到 localhost(我在 nginx 后面代理 Node 应用程序，但 Web 应用程序通过 Gulp 运行(，而远程 Node 应用程序位于端口 80 后面的不同域上。

任何线索都非常感谢:)

Answer 1

我遇到了完全相同的问题。在查看了一些 tcp 捕获后，看起来 nginx 甚至没有收到传入的升级标头，因此它没有将其传递给 node.js。(我确实怀疑Cloudflare这样做了，但我不确定(。

我解决这个问题的方式真的很笨拙，但它确实工作得很好。基本上我修改了nginx配置以查找Sec-Websocket-Key标头，当它找到时，它将升级标头设置为websocket，并将连接标头设置为upgrade。这行得通。

配置示例：

map $http_sec_websocket_key $upgr {
    ""      "";           # If the Sec-Websocket-Key header is empty, send no upgrade header
    default "websocket";  # If the header is present, set Upgrade to "websocket"
}
map $http_sec_websocket_key $conn {
    ""      $http_connection;  # If no Sec-Websocket-Key header exists, set $conn to the incoming Connection header
    default "upgrade";         # Otherwise, set $conn to upgrade
}

您必须在server {}块之前定义这两个地图块。

现在，在您的 socket.io 位置添加以下 2 行：

proxy_set_header Upgrade $upgr;
proxy_set_header Connection $conn;

这会将"升级"和"连接"标头设置为映射的值。

我希望这也是您问题的解决方案。