13 KiB

Raw Permalink Blame History

keepalived高可用集群

注意：centos7和9的配置完全一样

负载均衡 lb集群 load balance

流量分发

高可用 ha集群 high availability

主要是给服务器做冗余

keepalive 持久连接保持存活

keepalived 高可用软件名称

红帽有自己的高可用集群套件：RHCS

keepalived介绍

keepalived是集群管理中保证集群高可用的一个服务软件，其功能类似于heartbeat，用来防止单点故障。

master vip backup vip

脑裂问题：

backup vip

master vip

解决：stonith shoot the other node in the head 爆头

重启keepalived 关闭keepalived服务

keepalived工作原理

keepalived是以VRRP协议为实现基础的，是实现路由器高可用的协议，VRRP全称Virtual Router Redundancy Protocol，即虚拟路由冗余协议。

面试题(keepalived的工作原理) 将N台提供相同功能的服务器组成一个服务器组，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（该服务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

广播组播单播

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。

keepalived部署

++++++++++++

+---------------------------> + Client + 192.168.122.1/24 (真实机做客户端) | ++++++++++++ | | VIP eth0:1 192.168.122.254/24 | | | | ++++++++++++++++ +++++++++++++++++ | + Director master + + Director backup + | ++++++++++++++++ +++++++++++++++++ | DIP eth0 192.168.122.10/24 DIP eth0 192.168.122.20/24 | || |_______ ____________________ |_________________________________________ | | | |

++++++++++++++++ ++++++++++++++++

+ Real Server A + + Real Server B +

++++++++++++++++ ++++++++++++++++

eth0 192.168.122.30/24 eth0 192.168.122.40/24

分别在Director master、Director backup 上部署浮动资源（VIP IPVS策略） 测试2个Director在DR模式下都工作正常。测试完成后都撤掉浮动资源。

关于dr网卡路由条目上下顺序问题

因为dr2是ens36在下面，ens33在上面，所以配置keepalived的时候就不行了，可尝试以下解决方案调整路由条目上下顺序（如果还不行就需要在开始规划的时候就要固定网卡固定ip规划好，其实物理服务器不会出现这个问题）

[root@server ~]# ip r l
default via 192.168.26.2 dev ens36 proto dhcp src 192.168.26.132 metric 100 
default via 192.168.26.2 dev ens33 proto dhcp src 192.168.26.131 metric 101 
192.168.26.0/24 dev ens36 proto kernel scope link src 192.168.26.132 metric 100 
192.168.26.0/24 dev ens33 proto kernel scope link src 192.168.26.131 metric 101 

[root@server ~]# ip link set ens36 down
[root@server ~]# ip r l
default via 192.168.26.2 dev ens33 proto dhcp src 192.168.26.131 metric 101 
192.168.26.0/24 dev ens33 proto kernel scope link src 192.168.26.131 metric 101 

[root@server ~]# ip link set ens36 up
[root@server ~]# ip r l
default via 192.168.26.2 dev ens33 proto dhcp src 192.168.26.131 metric 101 
default via 192.168.26.2 dev ens36 proto dhcp src 192.168.26.132 metric 102 
192.168.26.0/24 dev ens33 proto kernel scope link src 192.168.26.131 metric 101 
192.168.26.0/24 dev ens36 proto kernel scope link src 192.168.26.132 metric 102

1. 在master上安装配置Keepalived

# yum install keepalived -y

2. 修改配置文件(清空原有配置文件)

# cd /etc/keepalived/
# vim keepalived.conf

//全局配置 ! Configuration File for keepalived global_defs { notification_email { root@localhost } notification_email_from keepalived@localhost smtp_server 127.0.0.1 smtp_connect_timeout 30 router_id Director1 //两边是一样的 } //局部配置 vrrp_instance VI_1 { state MASTER //另外一台机器是BACKUP interface eth0 //心跳网卡 DIP那一块网卡 virtual_router_id 51 priority 50 //优先级 advert_int 1 //检查间隔，单位秒 authentication { auth_type PASS auth_pass 1111 } virtual_ipaddress { 192.168.122.254/24 dev eth0 //VIP随便自己定义只要是同一个网段就可以 } }

virtual_server 192.168.122.254 80 { //LVS 配置，VIP delay_loop 3 //服务论询的时间间隔 lb_algo rr //LVS 调度算法 lb_kind DR // LVS 集群模式 protocol TCP real_server 192.168.122.30 80 { weight 1 TCP_CHECK { connect_timeout 3 } } real_server 192.168.122.40 80 { weight 1 TCP_CHECK { connect_timeout 3 } } }

3.在backup上安装keepalived

# yum install keepalived -y

4.拷贝master上的keepalived.conf到backup上：

# scp keepalived.conf 192.168.122.20:/etc/keepalived/

5.拷贝后,修改配置文件 state BACKUP priority 100

6.两个Director上启动服务

# systemctl start keepalived

7.测试 1）观察lvs路由条目 2）观察vip地址在哪台机器上（有可能两台机器上都有，但是他好用没有问题即可） 3）客户端浏览器访问vip 4）关闭master上的keepalived服务，再次访问vip

实验过程总结

10.0.0.20 dr1 10.0.0.21 dr2 10.0.0.22 web1 rs1 10.0.0.23 web2 rs2

1. 设置rs的配置 1.1 安装web服务器 1.2 添加vip：10.0.0.40 lo 1.3 arp 1 2

2. dr的配置 2.1 两块网卡两台机器上的两块网卡名称必须一样 2.2 路由条目的顺序必须一样 2.3 区分谁当dip 谁的路由条目在上面谁当dip 2.4 dip要固定下来 20 21 ens33 2.5 vip ens37 :ip地址有或没有都行只要这个网卡是启动状态如果有Ip地址，这个Ip不能是vip 2.6 安装ipvsadm

3. 安装keepalived到两台dr上 修改keepalived.conf配置文件：设置vip 设置角色 master backup 设置realserver是谁 ip地址是多少

扩展实验-keepalived+mysql

mysql可以是以下3种情况

1、双主

2、mysql-cluster

3、gelara集群

项目环境 VIP 192.168.122.100 mysql1 192.168.122.10 mysql2 192.168.122.20 vip 主(keepalived) 主(keepalived）

实现过程概要

一、mysql 主主同步二、在两台mysql上安装keepalived 三、keepalived 主备配置文件四、mysql状态检测脚本/root/bin/keepalived_check_mysql.sh 五、测试及诊断注 keepalived之间使用vrrp组播方式通信使用的IP地址是224.0.0.18 ＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝ 实施步骤 一、mysql 主主同步 <略>

二、安装keepalived

两台
# yum install keepalived -y

三、keepalived 主备配置文件 主备置文件不同处有 state priority

192.168.122.10 Master配置

# vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs { router_id mysql1 //两边一样 }

vrrp_script check_run { //定义配置健康检查脚本的配置名称check_run script "/root/keepalived_check_mysql.sh" interval 5 //执行健康检查的时间间隔单位s }

vrrp_instance VI_1 { state MASTER interface eth0 virtual_router_id 88 priority 100 advert_int 1 //检查keepalived本身服务的时间间隔 authentication { auth_type PASS auth_pass 1111 }

track_script { check_run //名字必须和上面的脚本配置名称一致 }

virtual_ipaddress { 192.168.122.100 dev eth0 } }

192.168.122.20 Slave配置

# vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs { router_id mysql1 }

vrrp_script check_run { script "/root/keepalived_check_mysql.sh" interval 5 }

vrrp_instance VI_1 { state BACKUP interface eth0 virtual_router_id 88 priority 90 advert_int 1 authentication { auth_type PASS auth_pass 1111 }

track_script { check_run }

virtual_ipaddress { 192.168.122.100 } }

四、mysql状态检测脚本

/root/keepalived_check_mysql.sh（两台MySQL同样的脚本）

版本一：简单使用：
#!/bin/bash
/usr/bin/mysql -h ip -uroot -p123 -e "show status;" &>/dev/null
if [ $? -ne 0 ] ;then
        systemctl stop keepalived
fi
# 此处的ip是本地ip

版本二：检查多次
#  vim  /root/keepalived_check_mysql.sh
#!/bin/bash
MYSQL=/usr/local/mysql/bin/mysql
MYSQL_HOST=localhost
MYSQL_USER=root
MYSQL_PASSWORD=1111
CHECK_TIME=3

#mysql  is working MYSQL_OK is 1 , mysql down MYSQL_OK is 0
MYSQL_OK=1

check_mysql_helth (){
    $MYSQL -h $MYSQL_HOST -u $MYSQL_USER -p${MYSQL_PASSWORD} -e "show status;" &>/dev/null
    if [ $? -eq 0 ] ;then
        MYSQL_OK=1
    else
        MYSQL_OK=0
    fi
    return $MYSQL_OK
}

while [ $CHECK_TIME -ne 0 ]
do
        check_mysql_helth
        if [ $MYSQL_OK -eq 1 ] ; then
                exit 0
        fi

        if [ $MYSQL_OK -eq 0 ] &&  [ $CHECK_TIME -eq 1 ];then
                /etc/init.d/keepalived stop
                exit 1
        fi
        let CHECK_TIME--
        sleep 1
done

版本三：检查多次
#  vim  /root/keepalived_check_mysql.sh
#!/bin/bash
MYSQL=/usr/local/mysql/bin/mysql
MYSQL_HOST=localhost
MYSQL_USER=root
MYSQL_PASSWORD=1111
CHECK_TIME=3

#mysql  is working MYSQL_OK is 1 , mysql down MYSQL_OK is 0
MYSQL_OK=1

check_mysql_helth (){
    $MYSQL -h $MYSQL_HOST -u $MYSQL_USER -p${MYSQL_PASSWORD} -e "show status;" &>/dev/null
    if [ $? -eq 0 ] ;then
        MYSQL_OK=1
    else
        MYSQL_OK=0
    fi
    return $MYSQL_OK
}

while [ $CHECK_TIME -ne 0 ]
do
        check_mysql_helth
        if [ $MYSQL_OK -eq 1 ] ; then
                exit 0
        fi

        let CHECK_TIME--
        sleep 1
done
/etc/init.d/keepalived stop
exit 1

# chmod 755 /root/keepalived_check_mysql.sh

两边均启动keepalived

日志查看脚本是否被执行
# tail -f /var/log/messages
Jun 19 15:20:19 xen1 Keepalived_vrrp[6341]: Using LinkWatch kernel netlink reflector...
Jun 19 15:20:19 xen1 Keepalived_vrrp[6341]: VRRP sockpool: [ifindex(2), proto(112), fd(11,12)]
Jun 19 15:20:19 xen1 Keepalived_vrrp[6341]: VRRP_Script(check_run) succeeded

扩展实验-keepalived+nginx负载均衡器

# vim /etc/keepalived/keepalived.conf
global_defs {
   router_id nginx1
}

vrrp_script check_run {
   script "/keepalived_check_nginx_proxy.sh"
   interval 5
}

vrrp_instance VI_1 {
    state MASTER
    interface ens32
    virtual_router_id 88
    priority 90
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }

    track_script {
        check_run
    }

    virtual_ipaddress {
        192.168.26.39
    }
}

# cat /keepalived_check_nginx_proxy.sh 
curl 127.1 &>/dev/null
if [ $? -ne 0 ];then
	systemctl stop keepalived
fi

13 KiB Raw Permalink Blame History Unescape Escape