一、原理及简介:
1.1 Keepalived简介
Keepalived是Linux下一个轻量级别的高可用解决方案。Keepalived起初是为LVS设计的,专门用来监控集群系统中各个服务节点的状态,它根据TCP/IP参考模型的第三、第四层、第五层交换机制检测每个服务节点的状态,如果某个服务器节点出现异常,或者工作出现故障,Keepalived将检测到,并将出现的故障的服务器节点从集群系统中剔除,这些工作全部是自动完成的,不需要人工干涉,需要人工完成的只是修复出现故障的服务节点。
后来Keepalived又加入了VRRP的功能,VRRP(Virtual Router Redundancy Protocol,虚拟路由冗余协议)出现的目的是解决静态路由出现的单点故障问题,通过VRRP可以实现网络不间断稳定运行,实现高可用性,因此Keepalvied 一方面具有服务器状态检测和故障隔离功能,另外一方面也有HA cluster功能,下面介绍一下VRRP协议实现的过程。
上图是Keepalived的功能体系结构,大致分两层:用户空间(user space)和内核空间(kernel space)。
内核空间:主要包括IPVS(IP虚拟服务器,用于实现网络服务的负载均衡)和NETLINK(提供高级路由及其他相关的网络功能)两个部份。
用户空间:
- WatchDog:负载监控checkers和VRRP进程的状况
- VRRP Stack:负载负载均衡器之间的失败切换FailOver,如果只用一个负载均稀器,则VRRP不是必须的。
- Checkers:负责真实服务器的健康检查healthchecking,是keepalived最主要的功能。换言之,可以没有VRRP Stack,但健康检查healthchecking是一定要有的。
- IPVS wrapper:用户发送设定的规则到内核ipvs代码
- Netlink Reflector:用来设定vrrp的vip地址等。
1.2 VRRP协议和工作原理
VRRP可以将两台或者多台物理路由器设备虚拟成一个虚拟路由,这个虚拟路由器通过虚拟IP(一个或者多个)对外提供服务,而在虚拟路由器内部十多个物理路由器协同工作,同一时间只有一台物理路由器对外提供服务,这台物理路由设备被成为:主路由器(Master角色),一般情况下Master是由选举算法产生,它拥有对外服务的虚拟IP,提供各种网络功能,如:ARP请求,ICMP 数据转发等,而且其它的物理路由器不拥有对外的虚拟IP,也不提供对外网络功能,仅仅接收MASTER的VRRP状态通告信息,这些路由器被统称为“BACKUP的角色”,当主路由器失败时,处于BACKUP角色的备份路由器将重新进行选举,产生一个新的主路由器进入MASTER角色,继续提供对外服务,整个切换对用户来说是完全透明的。
每个虚拟路由器都有一个唯一的标识号,称为VRID,一个VRID与一组IP地址构成一个虚拟路由器,在VRRP协议中,所有的报文都是通过IP多播(一对多,一个Master对多个Backup)方式发送的,而在一个虚拟路由器中,只有处于Master角色的路由器会一直发送VRRP数据包,处于BACKUP角色的路由器只会接受Master角色发送过来的报文信息,用来监控Master运行状态,一般不会发生BACKUP抢占MASTER的情况,除非它的优先级更高,而当MASTER不可用时,BACKUP也就无法收到Master发过来的信息,于是就认定Master出现故障,接着多台BAKCUP就会进行选举,优先级最高的BACKUP将称为新的MASTER,这种选举角色切换非常之快,因而保证了服务的持续可用性。
1.3 LVS负载均衡模式 直接路由模式(DR)
直接路由,通过为请求报文重新封装一个MAC首部进行转发,源MAC是DIP所在的接口的MAC,目标MAC是某挑选出的RS的RIP所在接口的MAC地址;源IP/PORT,以及目标IP/PORT均保持不变。
Director和各RS都得配置使用VIP;
(1) 确保前端路由器将目标IP为VIP的请求报文发往Director;
(a) 在前端网关做静态绑定;
(b) 在RS上使用arptables;
(c) 在RS上修改内核参数以限制arp通告及应答级别;
限制响应级别:arp_ignore
0:默认值,表示可使用本地任意接口上配置的任意地址进行响应;
1: 仅在请求的目标IP配置在本地主机的接收到请求报文接口上时,才给予响应;
限制通告级别:arp_announce
0:默认值,把本机上的所有接口的所有信息向每个接口上的网络进行通告;
1:尽量避免向非直接连接网络进行通告;
2:必须避免向非本网络通告;
(2) RS的RIP可以使用私网地址,也可以是公网地址;RIP与DIP在同一IP网络;RIP的网关不能指向DIP,以确保响应报文不会经由Director;
(3) RS跟Director要在同一个物理网络;
(4) 请求报文要经由Director,但响应不能经由Director,而是由RS直接发往Client;
(5) 不支持端口映射
2.3 集群配置前提
(1) 各节点时间必须同步(chrony or ntp)
(2) 确保iptables及selinux不会成为阻碍
(3) 各节点之间可通过主机名互相通信(并非必须)
建议使用/etc/hosts文件实现
(4) 确保各节点的用于集群服务的接口支持MULTICAST通信
打开或关闭支持组播通信 on|off
[root@Director1 ~]# ip link set multicast on dev ens33 //打开组播功能,默认为开启状态 [root@Director1 ~]# ifconfig | grep MULTICAST ens33: flags=4163MULTICAST> mtu 1500 ens33:0: flags=4163 MULTICAST> mtu 1500 virbr0: flags=4099 mtu 1500
三、前端调度服务器配置
(1)安装keepalived
[root@Director1 ~]# yum -y install keepalived //安装keepalived [root@Director1 ~]# cd /etc/keepalived/ [root@Director1 keepalived]# cp keepalived.conf{,.bak} //复制一份配置文件
(2)主配置文件(Director1)
[root@Director1 keepalived]# vim /etc/keepalived/keepalived.conf //编辑配置文件 ! Configuration File for keepalived ##################全局配置########################## global_defs { notification_email { root@localhost 371304@qq.com //设置报警邮件单个或多个地址 } notification_email_from keepalived@localhost //keepalived报警邮件,发件人 smtp_server 127.0.0.1 //邮件服务器地址 smtp_connect_timeout 30 //邮件服务器超时时间 router_id Direcotr1 //路由ID两台机器不能相同 vrrp_mcast_group4 224.0.0.18 //多播地址,范围224.0.0.0~239.255.255.255
vrrp_skip_check_adv_addr vrrp_strict vrrp_garp_interval 0 vrrp_gna_interval 0 } #################keepalived配置##################### vrrp_instance VI_1 { state MASTER //keepalived角色,MASTER和BACKUP interface ens33 //指定HA监测网络的接口
virtual_router_id 99 //虚拟路由ID号,0~255
,MASTER和BACKUP必须是一致的
priority 100 //定义优先级,数字越大 优先级越高 MASTER的优先级高于BACKUP优先级 advert_int 1 //设定MASTER与BACKUP负载均衡器之间同步检查的时间间隔,单位是秒
authentication { //设置验证类型和密码
auth_type PASS //设置验证类型,主要有PASS和AH两种
auth_pass QeiVWCxs //设置验证密码,在同一个vrrp_instance下,MASTER与BACKUP必须使用相同的密码
} //密码生成openssl rand -base64 6取前8位即可 virtual_ipaddress { //设置虚拟IP地址,可以设置多个虚拟IP地址,每行一个
192.168.10.99/24 dev ens33 label ens33:0 } notify_master "/etc/keepalived/notify.sh master" //调用通知脚本 notify_backup "/etc/keepalived/notify.sh backup" notify_fault "/etc/keepalived/notify.sh fault" } ##################LVS配置##############
#添加虚拟服务器 #相当于 ipvsadm -A -t 192.168.10.99:80 -s wrr virtual_server 192.168.10.99 80 { //设置虚拟服务器,需要指定虚拟IP地址和服务端口,IP与端口之间用空格隔开
delay_loop 1 //设置运行情况检查时间,单位是秒
lb_algo wrr //设置负载调度算法
lb_kind DR //设置LVS实现负载均衡的机制,有NAT、TUN、DR三个模式可选
protocol TCP //指定转发协议类型,有TCP和UDP两种
sorry_server 127.0.0.1 80 //当后端服务器全部脱机时,访问即调用本地 real_server 192.168.10.61 80 { //配置服务节点1,需要指定real server的真实IP地址和端口,IP与端口之间用空格隔开
weight 1 //配置服务节点的权值
HTTP_GET { //http进行检测 url { path /index.html status_code 200 } connect_timeout 3 //表示3秒无响应超时
nb_get_retry 3 //表示重试次数
delay_before_retry 2 //表示重试间隔
} } real_server 192.168.10.62 80 { //配置服务节点2
weight 1 HTTP_GET { url { path /index.html status_code 200 } connect_timeout 3 nb_get_retry 3 delay_before_retry 2 } } }
(3)备配置文件(Director2)
[root@Director2 ~]# cat /etc/keepalived/keepalived.conf ! Configuration File for keepalived global_defs { notification_email { root@localhost 371304@qq.com } notification_email_from keepalived@localhost smtp_server 127.0.0.1 smtp_connect_timeout 30 router_id Director2 //与主节点不一致 vrrp_mcast_group4 224.0.0.18 vrrp_skip_check_adv_addr vrrp_strict vrrp_garp_interval 0 vrrp_gna_interval 0 } vrrp_instance VI_1 { state BACKUP //状态为BACKUP interface ens33 virtual_router_id 99 //必须保持一致 priority 96 //优先级低于Master advert_int 1 authentication { auth_type PASS auth_pass QeiVWCxs } virtual_ipaddress { 192.168.10.99/24 dev ens33 label ens33:0 } notify_master "/etc/keepalived/notify.sh master" notify_backup "/etc/keepalived/notify.sh backup" notify_fault "/etc/keepalived/notify.sh fault" } virtual_server 192.168.10.99 80 { delay_loop 1 lb_algo wrr lb_kind DR protocol TCP sorry_server 127.0.0.1 80 real_server 192.168.10.61 80 { weight 1 HTTP_GET { url { path /index.html status_code 200 } connect_timeout 3 nb_get_retry 3 delay_before_retry 2 } } real_server 192.168.10.62 80 { weight 1 HTTP_GET { url { path /index.html status_code 200 } connect_timeout 3 nb_get_retry 3 delay_before_retry 2 } } }
(4)邮件通知脚本
当双主高可用集群主备切换时可通过邮件通知管理员,此时在配置文件中可自动调用实现编辑好的脚本
[root@Director1 keepalived]# vim /etc/keepalived/notify.sh //编辑脚本 #!/bin/bash # contact='root@localhost' notify() { local mailsubject="$(hostname) to be $1, vip floating" local mailbody="$(date +'%F %T'): vrrp transition, $(hostname) changed to be $1" echo "$mailbody" | mail -s "$mailsubject" $contact } case $1 in master) notify master ;; backup) notify backup ;; fault) notify fault ;; *) echo "Usage: $(basename $0) {master|backup|fault}" exit 1 ;; esac [root@Director1 keepalived]# chmod a+x /etc/keepalived/notify.sh //赋予执行权限 [root@Director1 keepalived]# scp -p /etc/keepalived/notify.sh Director2:/etc/keepalived/ //拷贝到Director2服务上的/etc/keepalived目录下 root@director2's password: notify.sh 100% 408 22.3KB/s 00:00 ~
(5)两节点安装httpd软件及配置默认首页
[root@Director1 ~]# yum -y install httpd [root@Director2 ~]# yum -y install httpd [root@Director1 ~]# echo 'Sorry!!!Server maintenance is in process. Please wait...
'>/var/www/html/index.html [root@Director2 ~]# echo 'Sorry!!!Server maintenance is in process. Please wait...
'>/var/www/html/index.html [root@Director1 ~]# systemctl start httpd.service;ssh Director2 'systemctl start httpd.service' root@director2's password: [root@Director1 ~]# systemctl enable httpd.service;ssh Director2 'systemctl enable httpd.service' //两节点设置开机自启动 Created symlink from /etc/systemd/system/multi-user.target.wants/httpd.service to /usr/lib/systemd/system/httpd.service. root@director2's password: Created symlink from /etc/systemd/system/multi-user.target.wants/httpd.service to /usr/lib/systemd/system/httpd.service. [root@Director1 ~]# curl 192.168.10.7 //测试主节点Sorry!!!Server maintenance is in process. Please wait...
[root@Director1 ~]# curl 192.168.10.8 //测试另一节点Sorry!!!Server maintenance is in process. Please wait...
(6)两节点安装ipvsadmin软件
[root@Director1 ~]# yum -y install ipvsadm [root@Director2~]# yum -y install ipvsadm [root@Director1 ~]# rpm -qa ipvsadm ipvsadm-1.27-7.el7.x86_64
四、后端RS服务器的配置
(1)后端真实服务器两节点安装httpd软件及配置默认首页
[root@RS1 ~]# yum -y install httpd [root@RS2 ~]# yum -y install httpd [root@RS1 ~]# echo "RS1:192.168.10.61
" > /var/www/html/index.html [root@RS2 ~]# echo "RS2:192.168.10.62
" > /var/www/html/index.html [root@RS1 ~]# /etc/init.d/httpd start;ssh RS2 '/etc/init.d/httpd start' Starting httpd: root@rs2's password: Starting httpd:
[root@RS1 ~]# chkconfig httpd on;ssh RS2 'chkconfig httpd on'
root@rs2's password: [root@Director1 ~]# curl 192.168.10.61RS1:192.168.10.61
[root@Director1 ~]# curl 192.168.10.62RS1:192.168.10.62
(2)修改RS内核参数可以通过脚本实现
a、 脚本编写如下:
#!/bin/bash #Desc: LVS_DR Real Server in the lo configuration VIP #Author:tony QQ:317104 #Date:2017-7-10 vip=192.168.10.99 mask='255.255.255.255' case $1 in start) echo 1 > /proc/sys/net/ipv4/conf/all/arp_ignore echo 1 > /proc/sys/net/ipv4/conf/lo/arp_ignore echo 2 > /proc/sys/net/ipv4/conf/all/arp_announce echo 2 > /proc/sys/net/ipv4/conf/lo/arp_announce ifconfig lo:0 $vip netmask $mask broadcast $vip up route add -host $vip dev lo:0 ;; stop) ifconfig lo:0 down echo 0 > /proc/sys/net/ipv4/conf/all/arp_ignore echo 0 > /proc/sys/net/ipv4/conf/lo/arp_ignore echo 0 > /proc/sys/net/ipv4/conf/all/arp_announce echo 0 > /proc/sys/net/ipv4/conf/lo/arp_announce ;; *) echo "Usage $(basename $0) start|stop" exit 1 ;; esac
b、将脚本复制到RS2目录下,并两台执行该脚本
[root@RS1 script]# chmod a+x /usr/src/script/lvs_DR.sh //赋予执行权限 [root@RS1 script]# cd /usr/src/script/ [root@RS1 script]# ./lvs_DR.sh start //执行脚本 [root@RS1 script]# scp -p lvs_DR.sh rs2:/usr/src/script/ root@rs2's password: lvs_DR.sh 100% 743 0.7KB/s 00:00 [root@RS2 ~]# /usr/src/script/lvs_DR.sh start
(3)查看后端两节点的VIP是否配置上,及相应的http的服务器已经起来
[root@RS1 script]# ip -4 addr 1: lo:mtu 65536 qdisc noqueue state UNKNOWN inet 127.0.0.1/8 scope host lo inet 192.168.10.99/32 brd 192.168.10.99 scope global lo:0 2: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000 inet 192.168.10.61/24 brd 192.168.10.255 scope global eth0 [root@RS1 script]# ss -tnl | grep 80 LISTEN 0 128 :::80 :::* [root@RS2 ~]# ip -4 addr 1: lo: mtu 65536 qdisc noqueue state UNKNOWN inet 127.0.0.1/8 scope host lo inet 192.168.10.99/32 brd 192.168.10.99 scope global lo:0 2: eth0: mtu 1500 qdisc pfifo_fast state UP qlen 1000 inet 192.168.10.62/24 brd 192.168.10.255 scope global eth0 [root@RS2 script]# ss -tnl | grep 80 LISTEN 0 128 :::80 :::*
五、测试
(1)开启Director2的keepalived服务
(2)查看ipvsadmin生成列表
[root@Director2 ~]# ipvsadm -L -n IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 192.168.10.99:80 wrr -> 192.168.10.61:80 Route 1 0 0 -> 192.168.10.62:80 Route 1 0 0
(3)测试访问VIP
[root@Director1 ~]# for i in {1..10};do curl http://www.contoso.com/index.html;doneRS1:192.168.10.61
RS2:192.168.10.62
RS1:192.168.10.61
RS2:192.168.10.62
RS1:192.168.10.61
RS2:192.168.10.62
RS1:192.168.10.61
RS2:192.168.10.62
RS1:192.168.10.61
RS2:192.168.10.62
(4)开启Director1的keepalived服务,查看及状态
[root@Director1 ~]# systemctl start keepalived.service //开启keepalived服务 [root@Director1 ~]# systemctl enable keepalived.service
[root@Director2 ~]# systemctl enable keepalived.service
(5)停掉后端两节点的http服务,查看LVS列表
[root@RS1 script]# /etc/init.d/httpd stop Stopping httpd: [ OK ] [root@Director1 ~]# ipvsadm -L -n //自动移除RS1节点 IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 192.168.10.99:80 wrr -> 192.168.10.62:80 Route 1 0 0 You have new mail in /var/spool/mail/root [root@RS2 ~]# /etc/init.d/httpd stop Stopping httpd: [ OK ] [root@Director1 ~]# ipvsadm -L -n IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 192.168.10.99:80 wrr -> 127.0.0.1:80 Route 1 0 0
(7)、启动RS2节点,再访问
[root@RS2 ~]# /etc/init.d/httpd start Starting httpd: [ OK ]
(8)、启动RS1节点,再访问,,LVS规则又正常了
[root@RS1 script]# /etc/init.d/httpd start Starting httpd: [ OK ] [root@Director1 ~]# ipvsadm -L -n IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 192.168.10.99:80 wrr -> 192.168.10.61:80 Route 1 0 2 -> 192.168.10.62:80 Route 1 2 0
(9)、停掉Director1的keepavlied服务,看是否节点转移成功
[root@Director1 ~]# systemctl stop keepalived.service //停掉keepalived服务 [root@Director1 ~]# ipvsadm -L -n IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn [root@Director2 ~]# ip -4 addr //查看Director2VIP地址是否绑定成功 1: lo:mtu 65536 qdisc noqueue state UNKNOWN qlen 1 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever 2: ens33: mtu 1500 qdisc pfifo_fast state UP qlen 1000 inet 192.168.10.8/24 brd 192.168.10.255 scope global ens33 valid_lft forever preferred_lft forever inet 192.168.10.99/24 scope global secondary ens33:0 valid_lft forever preferred_lft forever 3: virbr0: mtu 1500 qdisc noqueue state DOWN qlen 1000 inet 192.168.122.1/24 brd 192.168.122.255 scope global virbr0 valid_lft forever preferred_lft forever You have new mail in /var/spool/mail/root [root@Director2 ~]# ipvsadm -L -n //查看规则在Direcotr2上 IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 192.168.10.99:80 wrr -> 192.168.10.61:80 Route 1 0 0 -> 192.168.10.62:80 Route 1 0 0
(10)、检查邮件是否正常
(11)、使用抓包工具tcpdump查看VIP地址漂移情况,
[root@Director1 ~]# tcpdump -i ens33 -nn host 224.0.0.18
[root@Director1 ~]# tcpdump -i ens33 -vvv host 224.0.0.18
*****停掉Director1的keepalived服务之后,我们发现VIP瞬间转移至Director2上 *****