9.9 KiB
高可用集群
作者:行癫(盗版必究)
一:Keepalived简介
1.简介
keepalived是集群管理中保证集群高可用(HA)的一个服务软件,其功能类似于heartbeat,用来防止单点故障。
2.工作原理
keepalived是以VRRP协议为实现基础的,当backup收不到vrrp包时就认为master宕掉了,这时就需要根据VRRP的优先级来选举一个backup当master。这样我们就可以保证集群的高可用
keepalived是以VRRP协议为实现基础的,VRRP全称Virtual Router Redundancy Protocol,即虚拟路由冗余协议
虚拟路由冗余协议,可以认为是实现路由器高可用的协议,即将N台提供相同功能的路由器组成一个路由器组,这个组里面有一个master和多个backup,master上面有一个对外提供服务的vip(该路由器所在局域网内其他机器的默认路由为该vip),master会发组播,当backup收不到vrrp包时就认为master宕掉了,这时就需要根据VRRP的优先级来选举一个backup当master。这样的话就可以保证路由器的高可用了
keepalived主要有三个模块,分别是core、check和vrrp。core模块为keepalived的核心,负责主进程的启动、维护以及全局配置文件的加载和解析。check负责健康检查,包括常见的各种检查方式。vrrp模块是来实现VRRP协议的
如何判断谁是master:看vip,vip在谁上谁就是master
3.什么是脑裂
脑裂(split-brain):指在一个高可用(HA)系统中,当联系着的两个节点断开联系时,本来为一个整体的系统,分裂为两个独立节点,这时两个节点开始争抢共享资源,结果会导致系统混乱,数据损坏
对于无状态服务的HA,无所谓脑裂不脑裂;但对有状态服务(比如MySQL)的HA,必须要严格防止脑裂
究竟是有状态服务,还是无状态服务,其判断依据——两个来自相同发起者的请求在服务器端是否具备上下文关系
脑裂:backup强资源,master不认为自己会死,他俩抢着为客户端服务
解决方案:shoot the other in the head 爆头 master
注意:
在商城里购买一件商品。需要经过放入购物车、确认订单、付款等多个步骤。由于HTTP协议本身是无状态的,所以为了实现有状态服务,就需要通过一些额外的方案。比如最常见的session,将用户挑选的商品(购物车),保存到session中,当付款的时候,再从购物车里取出商品信息
二:LVS+Keepalived
1.环境准备
2.项目部署
主/备调度器安装软件
[root@lvs-keepalived-master ~]# yum -y install ipvsadm keepalived
[root@lvs-keepalived-slave ~]# yum -y install ipvsadm keepalived
主/备调度器配置
lvs-master:
[root@ha-proxy-master ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id lvs-keepalived-master #辅助改为lvs-backup
}
vrrp_instance VI_1 {
state MASTER
interface ens33 #VIP绑定接口
virtual_router_id 80 #VRID 同一组集群,主备一致
priority 100 #本节点优先级,辅助改为50
advert_int 1 #检查间隔,默认为1s
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.246.160/24
}
}
virtual_server 192.168.246.160 80 { #LVS配置
delay_loop 3
lb_algo rr #LVS调度算法
lb_kind DR #LVS集群模式(路由模式)
nat_mask 255.255.255.0
protocol TCP #健康检查使用的协议
real_server 192.168.246.162 80 {
weight 1
inhibit_on_failure #当该节点失败时,把权重设置为0,而不是从IPVS中删除
TCP_CHECK { #健康检查
connect_port 80 #检查的端口
connect_timeout 3 #连接超时的时间
}
}
real_server 192.168.246.163 80 {
weight 1
inhibit_on_failure
TCP_CHECK {
connect_timeout 3
connect_port 80
}
}
}
[root@lvs-keepalived-slave ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id lvs-keepalived-slave
}
vrrp_instance VI_1 {
state BACKUP
interface ens33
nopreempt #不抢占资源
virtual_router_id 80
priority 50
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.246.160/24
}
}
virtual_server 192.168.246.160 80 {
delay_loop 3
lb_algo rr
lb_kind DR
nat_mask 255.255.255.0
persistence_timeout 20
protocol TCP
real_server 192.168.246.162 80 {
weight 1
inhibit_on_failure
TCP_CHECK {
connect_port 80
connect_timeout 3
}
}
real_server 192.168.246.163 80 {
weight 1
inhibit_on_failure
TCP_CHECK {
connect_timeout 3
connect_port 80
}
}
}
3.LVS部署
[root@lvs-server ~]# ip addr add dev ens33 192.168.246.160/32 #设置VIP
[root@xingdian-cloud ~]# ipvsadm -S > /etc/sysconfig/ipvsadm
[root@xingdian-cloud ~]# systemctl start ipvsadm
[root@lvs-server ~]# ipvsadm -A -t 192.168.246.160:80 -s rr
[root@lvs-server ~]# ipvsadm -a -t 192.168.246.160:80 -r 192.168.246.163 -g
[root@lvs-server ~]# ipvsadm -a -t 192.168.246.160:80 -r 192.168.246.162 -g
[root@lvs-keepalived-master ~]# ipvsadm -Ln
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.246.160:80 rr persistent 20
-> 192.168.246.162:80 Route 1 0 0
-> 192.168.246.163:80 Route 0 0 0
4.启动KeepAlived
[root@lvs-keepalived-master ~]# systemctl start keepalived
[root@lvs-keepalived-master ~]# systemctl enable keepalived
5.RS配置
[root@test-nginx1 ~]# yum install -y nginx
[root@test-nginx2 ~]# yum install -y nginx
[root@test-nginx1 ~]# ip addr add dev lo 192.168.246.160/32
[root@test-nginx1 ~]# echo "net.ipv4.conf.all.arp_ignore = 1" >> /etc/sysctl.conf
[root@test-nginx1 ~]# echo 2 > /proc/sys/net/ipv4/conf/all/arp_announce
[root@test-nginx1 ~]# sysctl -p
[root@test-nginx1 ~]# echo "web1..." >> /usr/share/nginx/html/index.html
[root@test-nginx1 ~]# systemctl start nginx
三:LVS+Nginx
1.准备好web-server(两台都要做)
web-server-1:10.0.0.42
web-server-2:10.0.0.141
安装nginx,并且保证nginx正常运行
分别要在web-server上创建一个测试界面
echo "web-server-1" > /usr/share/nginx/html/index.html
echo "web-server-2" > /usr/share/nginx/html/index.html
检测web-server是否正常被访问
2.负载均衡的部署(两台都要做)
master:10.0.0.154
backup:10.0.0.27
master和backup都要做以下操作:
vim /etc/nginx/nginx.conf 添加以下内容
upstream xingdian {
server 10.0.0.42:80;
server 10.0.0.141:80;
}
vim /etc/nginx/conf.d/default.conf 修改一下内容
location / {
proxy_pass http://xingdian;
proxy_redirect default;
proxy_set_header Host $http_host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header REMOTE-HOST $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
}
保证nginx的负载均衡可用,客户端可以访问测试:
Keepalived实现调度器HA(vip我们直接写在配置文件中)
1. 主/备调度器安装软件(安装keepalived)
[root@master ~]# yum install -y keepalived
[root@backup ~]# yum install -y keepalived
[root@master ~]# cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf.bak (略)
[root@nginx-proxy-master ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id director1 #辅助改为director2
}
vrrp_instance VI_1 {
state MASTER #定义主还是备
interface ens33 #VIP绑定接口
virtual_router_id 80 #整个集群的调度器一致
priority 100 #back改为50
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
10.0.0.110/24
}
}
[root@backup ~]# cp /etc/keepalived/keepalived.conf /etc/keepalived/keepalived.conf.bak
[root@backup ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id directory2
}
vrrp_instance VI_1 { #实例名称,两台要保持相同
state BACKUP #设置为backup
interface ens33 #心跳网卡
nopreempt #设置到back上面,不抢占资源
virtual_router_id 80 #虚拟路由编号,主备要保持一致
priority 50 #辅助改为50
advert_int 1 #检查间隔,单位秒
authentication { 秘钥认证
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
10.0.0.110/24
}
}
3. 启动KeepAlived(主备均启动)
[root@master ~]# systemctl enable keepalived
[root@master ~]# systemctl start keepalived
四:健康检测
#!/bin/bash
#+检查nginx进程是否存在
counter=$(ps -C httpd --no-heading|wc -l) #此行有服务名
if [ "${counter}" = "0" ]; then
#尝试启动一次nginx,停止5秒后再次检测
service httpd start #启动服务
sleep 5
counter=$(ps -C httpd --no-heading|wc -l) #此行有服务名
if [ "${counter}" = "0" ]; then
#如果启动没成功,就杀掉keepalive触发主备切换
service keepalived stop
fi
fi