我在绑定到 unix 套接字的后端(带有 Puma 的 Rails)上有一个应用服务器,这是 nginx 配置的相关部分
location /live/ {
proxy_pass http://app; # match the name of upstream directive which is defined above
proxy_set_header Host $host;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_redirect off;
proxy_set_header Connection 'Upgrade';
proxy_http_version 1.1;
chunked_transfer_encoding off;
send_timeout 300;
proxy_send_timeout 300;
keepalive_timeout 7200;
}
SSE 事件来自 /live/,因此我调整了 ngix 配置以处理对该路由的所有请求。
问题是连接恰好在 60 秒后关闭。这是我在响应标头中看到的
Cache-Control:no-cache
Connection:keep-alive
Content-Type:text/event-stream
Date:Mon, 04 Nov 2013 13:41:52 GMT
Server:nginx
X-Content-Type-Options:nosniff
X-Frame-Options:SAMEORIGIN
X-Request-Id:7065a7bc-3450-4fe2-b60c-33dfa8d41951
X-Runtime:0.010852
X-UA-Compatible:chrome=1
X-XSS-Protection:1; mode=block
所以似乎 nginx 将初始响应设置为关闭。为什么keepalive_timeout在这里不起作用。
在 nginx error.log 我看到
2013/11/04 13:42:52 [error] 3689#0: *9 upstream timed out (110: Connection timed out) while reading upstream, client: ......., server: myapp.com, request: "GET /live/events HTTP/1.1", upstream: "http://unix:/home/ubuntu/myapp/shared/sockets/puma.sock:/live/events", host: "myapp.com"