HAProxy基础配置-haproxy常见的调度算法
作者：尹正杰
版权声明：原创作品，谢绝转载！否则将追究法律责任。
我们知道nginx有四种(rr,ip_hash,uri hash,lastconn)常用的调度算法，LVS有10种(常用的也就3-4种，而且生产环境种使用lvs的场景相对较少，毕竟大多数公司没有淘宝，天猫那个访问量)调度算法，而haproxy的调度算法相对nginx还是比较多的。
除此之外，haproxy还支持通过socket的方式实现服务器动态的上下限，而nginx和lvs却不支持该功能，接下来我们一起学习一下haproxy吧~
一.HAProxy调度算法需要使用balance指令来声明
balance:
指明对后端服务器的调度算法，配置在listen或backend
1>.编辑haproxy的配置文件(开启socket功能，便于下面测试动态调度算法与通过该socket和haproxy服务在不重启的情况下动态变更配置)
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# ll /yinzhengjie/softwares/haproxy/haproxy.sock
srw------- 1 root root 0 Jan 3 14:36 /yinzhengjie/softwares/haproxy/haproxy.sock
[[email protected] ~]#
[[email protected] ~]# ll /yinzhengjie/softwares/haproxy/haproxy.pid
-rw-r--r-- 1 root root 6 Jan 3 14:36 /yinzhengjie/softwares/haproxy/haproxy.pid
[[email protected] ~]#
2>.安装socat命令(下面在测试动态修改权重案例时会用到该工具)
[[email protected] ~]# yum -y install socat
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
epel/x86_64/metalink | 9.1 kB 00:00:00
* base: mirrors.aliyun.com
* epel: mirrors.yun-idc.com
* extras: mirror.bit.edu.cn
* updates: mirrors.aliyun.com
http://mirrors.ustc.edu.cn/centos/7.7.1908/os/x86_64/repodata/repomd.xml: [Errno 12] Timeout on http://mirrors.ustc.edu.cn/centos/7.7.1908/os/x86
_64/repodata/repomd.xml: (28, 'Operation too slow. Less than 1000 bytes/sec transferred the last 30 seconds')Trying other mirror.
base | 3.6 kB 00:00:00
extras | 2.9 kB 00:00:00
updates | 2.9 kB 00:00:00
Resolving Dependencies
--> Running transaction check
---> Package socat.x86_64 0:1.7.3.2-2.el7 will be installed
--> Finished Dependency Resolution
Dependencies Resolved
=================================================================================================================================================
Package Arch Version Repository Size
=================================================================================================================================================
Installing:
socat x86_64 1.7.3.2-2.el7 base 290 k
Transaction Summary
=================================================================================================================================================
Install 1 Package
Total download size: 290 k
Installed size: 1.1 M
Downloading packages:
socat-1.7.3.2-2.el7.x86_64.rpm | 290 kB 00:00:07
Running transaction check
Running transaction test
Transaction test succeeded
Running transaction
Installing : socat-1.7.3.2-2.el7.x86_64 1/1
Verifying : socat-1.7.3.2-2.el7.x86_64 1/1
Installed:
socat.x86_64 0:1.7.3.2-2.el7
Complete!
[[email protected] ~]#
[[email protected] ~]# echo "help" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #查看haproxy的socket使用的帮助信息 Unknown command. Please enter one of the following commands only : help : this message prompt : toggle interactive mode with prompt quit : disconnect show tls-keys [id|*]: show tls keys references or dump tls ticket keys when id specified set ssl tls-key [id|keyfile] <tlskey>: set the next TLS key for the <id> or <keyfile> listener to <tlskey> show errors : report last request and response errors for each proxy disable agent : disable agent checks (use 'set server' instead) disable health : disable health checks (use 'set server' instead) disable server : disable a server for maintenance (use 'set server' instead) enable agent : enable agent checks (use 'set server' instead) enable health : enable health checks (use 'set server' instead) enable server : enable a disabled server (use 'set server' instead) set maxconn server : change a server's maxconn setting set server : change a server's state, weight or address get weight : report a server's current weight set weight : change a server's weight (deprecated) show sess [id] : report the list of current sessions or dump this session shutdown session : kill a specific session shutdown sessions server : kill sessions on a server clear table : remove an entry from a table set table [id] : update or create a table entry's data show table [id]: report table usage stats or dump this table's contents clear counters : clear max statistics counters (add 'all' for all counters) show info : report information about the running process show stat : report counters for each proxy and server show schema json : report schema used for stats show startup-logs : report logs emitted during HAProxy startup show resolvers [id]: dumps counters from all resolvers section and associated name servers set maxconn global : change the per-process maxconn setting set rate-limit : change a rate limiting value set severity-output [none|number|string] : set presence of severity level in feedback information set timeout : change a timeout setting show env [var] : dump environment variables known to the process show cli sockets : dump list of cli sockets show fd [num] : dump list of file descriptors in use show activity : show per-thread activity stats (for support/developers) disable frontend : temporarily disable specific frontend enable frontend : re-enable specific frontend set maxconn frontend : change a frontend's maxconn setting show servers state [id]: dump volatile server information (for backend <id>) show backend : list backends in the current running config shutdown frontend : stop a specific frontend set dynamic-cookie-key backend : change a backend secret key for dynamic cookies enable dynamic-cookie backend : enable dynamic cookies on a specific backend disable dynamic-cookie backend : disable dynamic cookies on a specific backend show cache : show cache status add acl : add acl entry clear acl <id> : clear the content of this acl del acl : delete acl entry get acl : report the patterns matching a sample for an ACL show acl [id] : report available acls or dump an acl's contents add map : add map entry clear map <id> : clear the content of this map del map : delete map entry get map : report the keys and values matching a sample for a map set map : modify map entry show map [id] : report available maps or dump a map's contents show pools : report information about the memory pools usage [[email protected] ~]# [[email protected] ~]#
二.HAProxy的静态调度算法
静态算法:
按照事先定义好的规则轮询公平调度，不关心后端服务器的当前负载、链接数和相应速度等，且无法实时修改权重，只能重启后生效。
static-rr:
基于权重的轮询调度，不支持权重的运行时调整及后端服务器慢启动，其后端主机数量没有限制。
first:（生产环境使用场景较少）
根据服务器在列表中的位置，自上而下进行调度，但是其只会当第一台服务器的连接数达到上限，新请求才会分配给下一台服务，因此会忽略服务器的权重设置。
1>.测试static-rr静态调度算法
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #指定static-rr静态调度算法
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
balance static-rr
server web01 172.30.1.106:80 maxconn 3 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web02" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "set weight WEB_PORT_80/web01 3" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #我们发现static-rr静态调度算法不支持动态修改权重。
Backend is using a static LB algorithm and only accepts weights '0%' and '100%'.
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
2>.测试first静态调度算法(生产环境很少使用这种调度算法)
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #指定first调度算法
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
balance first
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 20`;do curl http://node102.yinzhengjie.org.cn/;done #很明显所有的请求都达到了web01的节点上。 <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> [[email protected] ~]# [[email protected] ~]#
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #使用first调度算法且指定服务器连接数上线为3
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
balance first
server web01 172.30.1.106:80 maxconn 3 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web02" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "set weight WEB_PORT_80/web01 3" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #很明显，first静态调度算法是支持动态修改权重的。
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
3 (initial 1)
[[email protected] ~]#
三.HAProxy的动态调度算法
动态算法：
基于后端服务器状态进行调度适当调整，比如优先调度至当前负载较低的服务器，且权重可以在haproxy运行时动态调整无需重启，若重启后，运行时配置的动态策略会被清空，依旧使用配置文件默认的配置哟~
roundrobin：
基于权重的轮询动态调度算法，支持权重的运行时调整，不等于lvs的rr，支持慢启动即新加的服务器会逐渐增加转发数，每个后端backend中最多支持4095个server，此为默认调度算法，server权重设置weight默认为1。
一般工作在四层做session共享用的最多的调度算法。
leastconn：
加权的最少连接的动态，支持权重的运行时调整和慢启动，即当前后端服务器连接最少的优先调度，比较适合长连接的场景使用，比如：LDAP,MySQL等场景。
一般工作在四层用于后端服务器MySQL,LDAP等场景。
1>.基于roundrobin动态调度算法调整权重案例
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #在配置文件中无需使用"balance"指定调度算法，因为默认就是"roundrobin"动态调度算法
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/;done #未修改权重之前，访问haproxy服务器 <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> [[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #默认的权重为1
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web02" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web03" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #由于我在配置文件中并没有定义web03节点，因此会返回"No such server"
No such server.
[[email protected] ~]#
[[email protected] ~]# echo "set weight WEB_PORT_80/web01 3" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #此处我将web01的权重动态修改为"3"
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #再次查看web01的权重果真变为3啦。
3 (initial 1)
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/;done #修改权重之后，再次访问haproxy服务器，很明显偏向"web01"节点 <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> [[email protected] ~]#
2>.服务器动态上下线案例(在haproxy 1.8多进程情况下貌似不支持，盼望官网早日修复该bug，但在haproxy 1.5确实是支持的)
[[email protected] ~]# echo "disable server WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #使节点下线，该方法在HAProxy1.5版本亲测是好使的，但是在HAProxy1.8却不太好使！希望官网早日修复吧
[[email protected] ~]#
[[email protected] ~]# echo "enable server WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #使节点上线，如果非要测试效果，可以使用yum方式安装haproxy1.5的版本来测试
[[email protected] ~]#
3>.leastconn调度算法配置案例
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #使用balance指定leastconn动态调度算法
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
balance leastconn
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 20`;do curl http://node102.yinzhengjie.org.cn/;done <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> [[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web02" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "set weight WEB_PORT_80/web01 3" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #很明显，leastconn动态调度算法也是支持动态修改权重的。
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
3 (initial 1)
[[email protected] ~]#
[[email protected] ~]#
四.source调度算法
source：
源地址hash，基于用户源地址hash并将请求转发到后端服务器，默认为静态(即取模方式)，但是可以通过hash-type支持的选项更改，后续同一个源地址请求将被转发至同一个后端web服务器，比较适用于后端服务器没有做session共享，但是还要实现session保持/缓存业务等场景。
map-based：
取模法，基于服务器权重的hash数组取模，该hash是静态的即不支持在线调整权重，不支持慢启动，其对后端服务器调度均衡，缺点是当服务器的总权重发生变化时，即有服务器上线或下线，都会因权重发生变化而导致调度结果整体改变。
consistent：
一致性哈希，该hash是动态的，支持在线调整权重，支持慢启动，优点在于当服务器的总权重发生变化时，对调度结果影响是局部的，不会引起大的变动。
source这种调度算法有一个致命缺点就是调度不均衡的问题，我们知道在一个公司局域网中所有的私网地址都会被NAT成一个公网IP地址，假设该公司有5万员工，那么该公司的五万员工的连接都会通过一个公网IP地址访问haproxy，由于这5万个主机都通过同一个公网IP地址访问haproxy，基于source调度算法会将这玩5万个主机的所有请求同时打到同一台后端web服务器，从而验证导致负载不均衡。
1>.source调度算法配置基于取模法(map-based,静态调度算法)案例
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #使用balance只当source调度算法
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
balance source
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 20`;do curl http://node102.yinzhengjie.org.cn/;done #使用haproxy本机测试访问，发现所有的请求都被调度到同一台服务器了 <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> <h1>node106.yinzhengjie.org.cnh1> [[email protected] ~]#
[[email protected] ~]# for _ in `seq 20`;do curl http://node102.yinzhengjie.org.cn/;done #使用172.30.1.107测试访问haproxy，发现所有的请求依旧被调度到同一台服务器了 <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> [[email protected] ~]# [[email protected] ~]#
[[email protected] ~]# for _ in `seq 20`;do curl http://node102.yinzhengjie.org.cn/;done #结果同上，所有的请求都被同一台web服务器响应了。 <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> <h2>node107.yinzhengjie.org.cnh1> [[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web02" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "set weight WEB_PORT_80/web01 3" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #很显然，source调度算法默认是静态算法，因为它不支持动态修改权重。
Backend is using a static LB algorithm and only accepts weights '0%' and '100%'.
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]#
2>.source调度算法配置基于一致性哈希(consistent，动态调度算法)案例
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #使用balance指定只当source调度算法并使用hash-type指定基于consistent(一致性哈希)，别忘记重启haproxy哟
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
balance source
hash-type consistent
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
五.uri(uniform resource identifier，统一资源标识符,是一个用于标识某一互联网资源名称的字符串)调度算法
uri：
基于对用户请求的uri做hash并将请求转发到后端指定服务器,适用于缓存业务场景。换句话说，同一个节点访问不同的uri可能会被调度到不同的后端服务器，但是不同的浏览器访问相同的uri则会被调度到同一台服务器，因为基于uri调度是在服务端完成的而非客户端。这种调度算法多用在类似CDN缓存场景。
和source调度算法类似，uri调度算法也支持取模法和一致性哈希，配置方式也一样，默认都是取模法，如下所示：
map-based：
取模法，基于服务器权重的hash数组取模，该hash是静态的即不支持在线调整权重，不支持慢启动，其对后端服务器调度均衡，缺点是当服务器的总权重发生变化时，即有服务器上线或下线，都会因权重发生变化而导致调度结果整体改变。
consistent：
一致性哈希，该hash是动态的，支持在线调整权重，支持慢启动，优点在于当服务器的总权重发生变化时，对调度结果影响是局部的，不会引起大的变动。
1>.uri调度算法使用一致性哈希(consistent)案例
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #使用balance指定haproxy调度算法为uri并使用hash-type指定为一致性哈希算法(别忘记重启haproxy)
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
balance uri
hash-type consistent
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app01/index.html;done #很明显，访问不同的uri会被调度到不同的节点
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app02/index.html;done
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app01/index.html;done #很明显，访问相同的uri无论是哪个客户端都被调度到相同的节点。
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app02/index.html;done
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
[[email protected] ~]#
[[email protected] ~]#
2>.uri调度算法配置基于取模法(map-based,静态调度算法)案例
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #使用balance指定haproxy调度算法为uri，无需使用hash-type，因为默认就是map-based(别忘记重启haproxy)
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
balance uri
#若不配置hash-type,则模式使用"hash-type map-based"
#hash-type consistent
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app01/index.html;done #客户端访问结果和上面一样
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app02/index.html;done
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
[[email protected] ~]#
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app01/index.html;done
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
172.30.1.107 app01
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app02/index.html;done
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
[[email protected] ~]#
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web02" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "set weight WEB_PORT_80/web01 3" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #静态调度算法不支持动态修改权重
Backend is using a static LB algorithm and only accepts weights '0%' and '100%'.
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]#
六.url_param调度算法(生产环境使用的较少)
url_param： 对用户请求的url中的<params>部分中的参数name作hash计算，并由服务器总权重相除以后派发至某挑出的服务器；通常用于追踪用户，以确保来自同一个用户的请求始终发往同一个Backend Server。这种调度算法也多用在CDN的缓存场景。
1>.实战案例
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg #使用balance指定url_param调度算法，需要指定一个参数名称哟(别忘记重启haproxy)
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
#基于user参数来调度
balance url_param user
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app02/index.html;done #我们在url不传参数的话默认调度类似于轮询。
172.30.1.106 app02
172.30.1.107 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.107 app02
172.30.1.106 app02
172.30.1.107 app02
172.30.1.106 app02
172.30.1.107 app02
172.30.1.107 app02
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app02/index.html?user=yinzhengjie;done #我们传参数后调度到同一台服务器了
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
172.30.1.106 app02
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app02/index.html?user=jason;done #我们传递的参数不一致发现可能会调度到不同的节点
172.30.1.107 app02
172.30.1.107 app02
172.30.1.107 app02
172.30.1.107 app02
172.30.1.107 app02
172.30.1.107 app02
172.30.1.107 app02
172.30.1.107 app02
172.30.1.107 app02
172.30.1.107 app02
[[email protected] ~]#
[[email protected] ~]# for _ in `seq 10`;do curl http://node102.yinzhengjie.org.cn/app02/index.html?name=jason;done #我们不适用服务端定义的user传参时发现并没有效果。
172.30.1.106 app02
172.30.1.106 app02
172.30.1.107 app02
172.30.1.106 app02
172.30.1.107 app02
172.30.1.106 app02
172.30.1.107 app02
172.30.1.107 app02
172.30.1.106 app02
172.30.1.107 app02
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web02" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
[[email protected] ~]# echo "set weight WEB_PORT_80/web01 3" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock #很显然，url_param调度算法也是静态的，并不支持动态修改权重。
Backend is using a static LB algorithm and only accepts weights '0%' and '100%'.
[[email protected] ~]#
[[email protected] ~]# echo "get weight WEB_PORT_80/web01" | socat stdio /yinzhengjie/softwares/haproxy/haproxy.sock
1 (initial 1)
[[email protected] ~]#
七.hdr调度算法(生产环境使用较少)
hdr(<name>)： 针对用户每个http头部(header)请求中的指定信息做hash，此处由<name>指定的http首部将会被取出并做hash计算，然后由服务器总权重相除以后派发至某挑出的服务器，假如无有效的值，则会被轮询调度。 和source调度算法类似，hdr调度算法也支持取模法和一致性哈希，配置方式也一样，默认都是取模法，如下所示： map-based： 取模法，基于服务器权重的hash数组取模，该hash是静态的即不支持在线调整权重，不支持慢启动，其对后端服务器调度均衡，缺点是当服务器的总权重发生变化时，即有服务器上线或下线，都会因权重发生变化而导致调度结果整体改变。 consistent： 一致性哈希，该hash是动态的，支持在线调整权重，支持慢启动，优点在于当服务器的总权重发生变化时，对调度结果影响是局部的，不会引起大的变动。
1>.实战案例
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
bind 172.30.1.102:80
mode http
#基于流览器客户端来调度
balance hdr(User-Agent)
hash-type consistent
server web01 172.30.1.106:80 check inter 3000 fall 3 rise 5
server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
[[email protected] ~]# curl -A 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/79.0.3945.88 Safari/537.36' http://node102.yinzhengjie.org.cn/ <h1>node106.yinzhengjie.org.cnh1> [[email protected] ~]# [[email protected] ~]# curl -A 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:47.0) Gecko/20100101 Firefox/47.0' http://node102.yinzhengjie.org.cn/ <h1>node106.yinzhengjie.org.cnh1> [[email protected] ~]# [[email protected] ~]# curl -A 'UCWEB7.0.2.37/28/999' http://node102.yinzhengjie.org.cn/ #这是模拟的UC浏览器 <h2>node107.yinzhengjie.org.cnh1> [[email protected] ~]# [[email protected] ~]# curl http://node102.yinzhengjie.org.cn/ #这里就没有使用任何浏览器模拟，而是直接使用curl命令工具访问。 <h1>node106.yinzhengjie.org.cnh1> [[email protected] ~]#
八.rdp-cookie调度算法(生产环境使用较少)
rdp-cookie:
对远程桌面的负载，使用cookie保持会话
和source调度算法类似，rdp-cookie调度算法也支持取模法和一致性哈希，配置方式也一样，默认都是取模法，如下所示：
map-based：
取模法，基于服务器权重的hash数组取模，该hash是静态的即不支持在线调整权重，不支持慢启动，其对后端服务器调度均衡，缺点是当服务器的总权重发生变化时，即有服务器上线或下线，都会因权重发生变化而导致调度结果整体改变，hash（o）modn 。
consistent：
一致性哈希，该hash是动态的，支持在线调整权重，支持慢启动，优点在于当服务器的总权重发生变化时，对调度结果影响是局部的，不会引起大的变动。
1>.修改haproxy支持window远程桌面功能
[[email protected] ~]# cat /etc/haproxy/haproxy.cfg
global
maxconn 100000
chroot /yinzhengjie/softwares/haproxy
#如果需要使用动态调度算法需要将socket功能打开
stats socket /yinzhengjie/softwares/haproxy/haproxy.sock mode 600 level admin
user haproxy
group haproxy
daemon
nbproc 2
cpu-map 1 0
cpu-map 2 1
nbthread 2
pidfile /yinzhengjie/softwares/haproxy/haproxy.pid
log 127.0.0.1 local5 info
defaults
option http-keep-alive
option forwardfor
option redispatch
option abortonclose
maxconn 100000
mode http
timeout connect 300000ms
timeout client 300000ms
timeout server 300000ms
listen stats
mode http
bind 0.0.0.0:9999
stats enable
log global
stats uri /haproxy-status
stats auth haadmin:q1w2e3r4ys
listen WEB_PORT_80
#注意，Windows的默认远程端口是3389哟，咱们haproxy监听端口也需要跟着变化，否则客户端还得显式输入端口才能访问。
bind 172.30.1.102:3389
#注意，如果代理的是windows远程桌面就不要使用http啦，一定要改为tcp模式
mode tcp
#基于流览器客户端来调度
balance rdp-cookie
hash-type consistent
#注意，windows默认的远程端口是3389(除非你自己手动修改过window远程服务的默认端口)，千万别写错了哈
server windows 172.30.1.254:3389 check inter 3000 fall 3 rise 5
#server web01 172.30.1.106: check inter 3000 fall 3 rise 5
#server web02 172.30.1.107:80 check inter 3000 fall 3 rise 5
[[email protected] ~]#
[[email protected] ~]# systemctl restart haproxy
[[email protected] ~]#
2>.确保window是开启了远程桌面的功能，如下图所示。
3>.使用宿主机远程haproxy服务器的3389服务，会弹出一个连接窗口，登录成功则会访问通过访问haproxy的端口从而远程了目标window主机哟，如下图所示。
九.算法总结
动态调度算法(tcp/http):
roundrobin
leastconn
静态调度算法(tcp/http)
static-rr
first
若指定hash-type的值为一致性哈希算法("consistent"),即动态调度算法,若不指定hash-type的值，则默认为取模法("map-based")，即静态调度算法。
source(支持tcp/http）
uri(支持http)
url_param(支持http)
hdr(支持http)
rdp-cookie(支持tcp)
如何选择haproxy工作在哪一层呢?
haproxy工作在第四层性能明显要高于haproxy工作在第七层，因为工作在第四层只涉及到数据报文的转发问题;而工作在第七层则需要对客户端的数据报文进行解封装操作。
关于HAProxy的四层与七层的区别博主推荐阅读:https://www.cnblogs.com/yinzhengjie/p/12127959.html.