nginx报告upstream timed out (10060: A connection attempt failed）但实际上服务收到了请求

Posted 2023-03-23

【中文标题】nginx报告upstream timed out (10060: A connection attempt failed）但实际上服务收到了请求

【英文标题】：nginx report upstream timed out (10060: A connection attempt failed）but actually the service received the request

【发布时间】：2021-11-18 19:52:36

【问题描述】：

操作系统：Windows 10

服务类型：Spring Boot搭建的HTTP服务器

我的场景是提供附件上传接口，附件经过base64编码后作为body参数上传到服务。它在大多数情况下都能正常工作。今天碰到一个请求，大小大概6M左右，每次nginx都会报“upstream timed out (10060: A connection attempt failed”。吹的是nginx的errorlog。

于是我用wireshark抓包，发现两个网络包之间有1分钟的间隔，可能是这个原因，但不知道怎么解释。有没有人遇到过同样的问题？

Wireshark 图片：

nginx错误日志：

2021/09/26 14:02:59 [debug] 24844#25224: *115 http upstream dummy handler
2021/09/26 14:02:59 [debug] 24844#25224: worker cycle
2021/09/26 14:02:59 [debug] 24844#25224: select event: fd:560 wr:0
2021/09/26 14:02:59 [debug] 24844#25224: select event: fd:608 wr:0
2021/09/26 14:02:59 [debug] 24844#25224: select event: fd:616 wr:0
2021/09/26 14:02:59 [debug] 24844#25224: select timer: 57
2021/09/26 14:02:59 [debug] 24844#25224: select ready 0
2021/09/26 14:02:59 [debug] 24844#25224: timer delta: 63
2021/09/26 14:02:59 [debug] 24844#25224: *115 event timer del: 616: 548607287
2021/09/26 14:02:59 [debug] 24844#25224: *115 http upstream request: "/xxx/yyy?"
2021/09/26 14:02:59 [debug] 24844#25224: *115 http upstream process upstream
2021/09/26 14:02:59 [error] 24844#25224: *115 upstream timed out (10060: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond) while reading upstream, client: 127.0.0.1, server: localhost, request: "POST /xxx/yyy HTTP/1.1", upstream: "http://127.0.0.1:9099/xxx/yyy", host: "localhost:80"
2021/09/26 14:02:59 [debug] 24844#25224: *115 http upstream exit: 00000000
2021/09/26 14:02:59 [debug] 24844#25224: *115 finalize http upstream request: 502
2021/09/26 14:02:59 [debug] 24844#25224: *115 finalize http proxy request
2021/09/26 14:02:59 [debug] 24844#25224: *115 free rr peer 1 0
2021/09/26 14:02:59 [debug] 24844#25224: *115 close http upstream connection: 616
2021/09/26 14:02:59 [debug] 24844#25224: *115 free: 02B7D598, unused: 88
2021/09/26 14:02:59 [debug] 24844#25224: *115 select del event fd:616 ev:768````

【参考方案1】：

增加proxy_read_timeout

proxy_read_timeout 300;

【讨论】：

谢谢，它成功了。但是花了60s才回复，我还是不明白为什么这么久。

不错，可能是上游服务器花了这么长时间，你能把这个标记为答案