mirror of
https://github.com/fatedier/frp.git
synced 2026-05-15 08:05:49 -06:00
[GH-ISSUE #714] 在阿里云添加了负载均衡后监听端口会导致云服务器上的FRP报错 #562
Labels
No labels
In Progress
WIP
WaitingForInfo
bug
doc
duplicate
easy
enhancement
future
help wanted
invalid
lifecycle/stale
need-issue-template
need-usage-help
no plan
proposal
pull-request
question
todo
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: github-starred/frp#562
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @wuji530 on GitHub (Apr 19, 2018).
Original GitHub issue: https://github.com/fatedier/frp/issues/714
What version of frp are you using (./frpc -v or ./frps -v)?
版本：服务端客户端均为 frp_0.16.1_windows_amd64
What operating system and processor architecture are you using (
go env)?系统：WIN2008 R2 企业版 SP1 x64位系统
1、在阿里云服务器开启FRPS 服务端程序，监听端口7000， 如果在阿里云负载均衡里也添加上7000端口，就会导致服务端FRPS报错：Accept new mux stream error: broken pipe
2、在本地局域网电脑启动客户端 启动FRPC 配置TCP 30001端口 配置如下，FRPC中服务端地址是直接连接阿里云服务器地址。
[common]
server_addr = 【此处为阿里云服务器IP地址】
server_port = 7000
[range:tcp_port]
type = tcp
local_ip = 127.0.0.1
local_port = 30001
remote_port = 30001
use_encryption = false
use_compression = false
负载均衡里未添加30001端口时可以正常,一添加上30001端口就会导致 客户端FRPC连接出错，错误内容如下：
2018/04/19 11:02:52 [E] [proxy.go:438] [tcp_port_0] connect to local service [12
7.0.0.1:30001] error: dial tcp 127.0.0.1:30001: connectex: No connection could b
e made because the target machine actively refused it.
停止负载均衡监听端口后又能正常连上FRPS服务端 所以确定是阿里云负载均衡会影响到FRPS服务端。
但是我客户端FRPC未启动时，如果云服务器上开启FRPS服务端,并且负载均衡添加30001端口后，FRPS未显示错误信息，而且natstat -an [查看端口也没有发现30001端口被占用，但是一旦开FRPC客户端进行连接就会报以上错误]
@wuji530 commented on GitHub (Apr 19, 2018):
另补充：我自己写的简单的服务端程序和客户端程序，在负载均衡中添加服务器程序的监听端口，服务端软件依然可以正常运行，客户端也能正常连接
@wuji530 commented on GitHub (Apr 19, 2018):
之后服务端更换Linux版，在Centos6.8 系统下 运行
如果负载均衡监听7000端口 也一样会导致frps程序报错
2018/04/19 13:01:41 [W] [service.go:247] Accept new mux stream error: broken pipe
@fatedier commented on GitHub (Apr 19, 2018):
根据你的描述，显然就是负载均衡器的问题了，你需要弄清楚这个组件的行为，以及对连接做了哪些控制或者修改。另外，frps 是有状态服务，只能部署一个，不能做负载均衡。
@wuji530 commented on GitHub (Apr 19, 2018):
我使用负载均衡的目的是因为负载均衡有5G的DDOS防护能力，另外采用负载均衡可以避免让云服务器IP暴露在外，价格也更加低廉 出故障后 重新添加负载均衡也能非常快速布置，并非要让FRPS服务器部署多个的目的。
目前看到的原因猜测是负载均衡添加端口后要检测端口状态，向服务器对应端口发送了某些检测指令，导致服务器端判断为端口被占用，而实际该端口并未占用
经过实际测试 我发现服务端开启端口后在负载均衡监听此端口会通过不断 连接进入、断开连接来探测端口是否正常开放，并未发送任何报文信息。 是否反复的连接断开来探测某端口会导致FRP以为端口被占用呢？
客户进入:100.97.199.129:144
客户离开:100.97.199.129:144
客户进入:100.97.198.131:54031
客户离开:100.97.198.131:54031
客户进入:100.109.154.130:12700
客户离开:100.109.154.130:12700
客户进入:100.109.154.3:469
客户离开:100.109.154.3:469
客户进入:100.97.198.2:31534
客户离开:100.97.198.2:31534
客户进入:100.97.199.2:53404
客户离开:100.97.199.2:53404
客户进入:100.109.153.128:59099
客户离开:100.109.153.128:59099
客户进入:100.109.153.2:15201
客户离开:100.109.153.2:15201
客户进入:100.97.198.131:7608
客户离开:100.97.198.131:7608
客户进入:100.97.199.129:64073
客户离开:100.97.199.129:64073
客户进入:100.97.198.2:5105
客户离开:100.97.198.2:5105
客户进入:100.109.154.130:16656
客户离开:100.109.154.130:16656
客户进入:100.97.199.2:7279
客户离开:100.97.199.2:7279
@wuji530 commented on GitHub (Apr 19, 2018):
自己做了个客户端模拟反复连接断开行为，发现确实是这个原因导致FRPS服务端和FRPC客户端报错
比如 FRPC客户端配置
[range:tcp_port]
type = tcp
local_ip = 127.0.0.1
local_port = 30005
remote_port = 30005
use_encryption = false
use_compression = false
我另外一个程序反复连接断开 FRPS服务端的 30005端口，就会导致FRPC客户端不断报错
2018/04/20 01:41:45 [E] [proxy.go:438] [tcp_port_0] connect to local service [
7.0.0.1:30005] error: dial tcp 127.0.0.1:30005: connectex: No connection could
e made because the target machine actively refused it.
2018/04/20 01:41:45 [E] [proxy.go:438] [tcp_port_0] connect to local service [
7.0.0.1:30005] error: dial tcp 127.0.0.1:30005: connectex: No connection could
e made because the target machine actively refused it.
2018/04/20 01:41:45 [E] [proxy.go:438] [tcp_port_0] connect to local service [
7.0.0.1:30005] error: dial tcp 127.0.0.1:30005: connectex: No connection could
e made because the target machine actively refused it.
2018/04/20 01:41:45 [E] [proxy.go:438] [tcp_port_0] connect to local service [
7.0.0.1:30005] error: dial tcp 127.0.0.1:30005: connectex: No connection could
e made because the target machine actively refused it.
2018/04/20 01:41:46 [E] [proxy.go:438] [tcp_port_0] connect to local service [
7.0.0.1:30005] error: dial tcp 127.0.0.1:30005: connectex: No connection could
e made because the target machine actively refused it.
2018/04/20 01:41:46 [E] [proxy.go:438] [tcp_port_0] connect to local service [
7.0.0.1:30005] error: dial tcp 127.0.0.1:30005: connectex: No connection could
e made because the target machine actively refused it.
2018/04/20 01:41:46 [E] [proxy.go:438] [tcp_port_0] connect to local service [
7.0.0.1:30005] error: dial tcp 127.0.0.1:30005: connectex: No connection could
e made because the target machine actively refused it.
2018/04/20 01:41:47 [E] [proxy.go:438] [tcp_port_0] connect to local service [
7.0.0.1:30005] error: dial tcp 127.0.0.1:30005: connectex: No connection could
e made because the target machine actively refused it.
@fatedier commented on GitHub (Apr 20, 2018):
127.0.0.1:30005是你的内网服务的端口，这个错误提示不是端口被占用了，而是连接不上，确认下这个服务是否正常监听在127.0.0.1:30005这个地址上。