求助:客户端偶发[E1008]Reached timeout
- Dominant language
- C++
- Stars
- 17.6k
- Forks
- 4.1k
- Avg merge
- 2d 12h
- Merged PRs (30d)
- 69
Description
**问题描述**
客户端偶发出现 `[E1008]Reached timeout=15000ms`,从 rpcz 抓到的trace看,服务端很快进行了response,但是不知道什么原因客户端一直没有收到,其中一个trace如下
```
09:38:09.066607 . 0 Requesting xxxxx baidu_std log_id=0 call_id=146673133403889 trace=d2a8ed1019254ce4 span=d2a8ed1019254ce4
09:38:09.066624 . 17 Requested(3581) [1]
09:38:24.066684 15. 60 [E1008]Reached timeout=15000ms @xxxx:8060
09:38:24.066710 . 26 Enter user's done
2024/04/22-09:38:09.067145 Received request(3581) from xxxx:35494 baidu_std log_id=0 trace=d2a8ed1019254ce4 span=d2a8ed1019254ce4
09:38:09.067176 . 31 Processing the request in a new bthread
09:38:09.067241 . 65 Enter xxxxx
09:38:09.067395 . 154 Leave xxxxx
09:38:09.067407 . 12 Responded(32)
```
按照 [server_debugging](https://github.com/apache/brpc/blob/master/docs/cn/server_debugging.md)对客户端和服务端进行了排查
* bthread_worker_usage 很低 (但依然尝试增加 brpc_concurrency,没有效果)
* 机器的CPU、IO和网络流量都不高
请问下有没有其它可能的原因或排查手段呢
**Versions (各种版本)**
brpc: 1.3.0
Contributor guide
Research direction
Start by reading docs/cn/server_debugging.md and comparing the client and server rpcz traces shown for brpc 1.3.0, including the bthread_worker_usage observation. The issue names no source entry point or test; completion would require identifying a reproducible cause or documenting a concrete diagnostic path.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- backend-api-design, networking
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100