haproxy implements tcp proxy2016-07-29 09:44:58 haproxy implements tcp proxy

haproxy implements tcp proxy

July 29, 2016 09:44:58

Readings: 3589

Introduction

In most cases, haproxy is an http (layer seven) proxy, such as apache, tomcat, etc. Let's talk about the tcp (four layer) proxy of haproxy, which can be used in ssh, mysql, mongodb and other occasions.

need

ip	application	Role
10.10.10.15	haproxy	tcp proxy
10.10.10.16	mongodb	master
10.10.10.17	mongodb	slave

haproxy acts as a proxy for mongodb to achieve master-slave read-write separation.

Install

 unzip haproxy-1.6.4.zip 
 cd haproxy-1.6.4
 make TARGET=linux26 PREFIX=/usr/local/haproxy 
 make install PREFIX=/usr/local/haproxy

• 1
• 2
• 3
• 4

configure

1. Add haproxy user and create related directories

useradd haproxy
mkdir -p /usr/local/haproxy/{etc,logs}

• 1
• 2

2. Configure log output

vim /etc/rsyslog.conf
#haproxy
$ModLoad imudp
$UDPServerRun 514
local3.* /usr/local/haproxy/logs/haproxy.log

service rsyslog restart

• 1
• 2
• 3
• 4
• 5
• 6
• 7

3. Configuration file

#
# demo config for Proxy mode
# 

global
        maxconn         20000 #最大连接数
    ulimit-n        204800  #ulimit的数量限制
        log             127.0.0.1 local3
        user             haproxy 
        group            haproxy
        chroot          /var/empty
    nbproc      4 #启动后运行的进程数量
        daemon #以后台形式运行haproxy
    pidfile     /var/run/haproxy.pid
defaults
    log global  
    mode tcp  #代理的级别（7层http，4层tcp）
    retries 3 #3次连接失败认为服务不可用，也可以在后面设置
    timeout connect 5s #连接超时
    timeout client 30s #客户端超时
    timeout server 30s #服务器端超时
    option redispatch  
    option          nolinger
    no option dontlognull
    option      tcplog
    option log-separate-errors


listen admin_stats  #监控页面设置
    bind 0.0.0.0:26000
    bind-process 1
    mode http
    log 127.0.0.1 local3 err
    stats refresh 30s ##每隔30秒自动刷新监控页面
    stats uri /haproxy_status
    stats realm welcome login\ Haproxy
    stats auth admin:admin
    stats hide-version
    stats admin if TRUE

frontend mongodb_read   #读代理
    bind        10.10.10.15:27000
    default_backend db_read


frontend mongodb_write  #写代理
    bind        10.10.10.15:27001
    default_backend db_write

backend db_read  #master slave都可以读
    balance roundrobin #负载均衡的方式，roundrobin是轮询
    server db1 10.10.10.16:27017 check inter 1500 rise 3 fall 3 weight 2
    #check inter 1500是检测心跳频率,rise 3是3次正确认为服务器可用，fall 3是3次失败认为服务器不可用，weight代表权重 
    server db2 10.10.10.17:27017 check inter 1500 rise 3 fall 3 weight 2

backend db_write #master 支持写
    server db1 10.10.10.16:27017

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30
• 31
• 32
• 33
• 34
• 35
• 36
• 37
• 38
• 39
• 40
• 41
• 42
• 43
• 44
• 45
• 46
• 47
• 48
• 49
• 50
• 51
• 52
• 53
• 54
• 55
• 56
• 57

4. Start

/usr/local/haproxy/sbin/haproxy -f /usr/local/haproxy/etc/haproxy.cfg

• 1

Log format analysis