准备工作
本例中,使用keepalive+nginx实现高可用和负载均衡。
资源要求
服务器
至少2台,如下:
VIP | IP | 主机名 | nginx端口 |
10.120.26.111 | 10.120.26.90 | status | 80 |
10.120.26.111 | 10.120.26.91 | LB2 | 80 |
IP地址资源
除了服务器IP外,还需要额外提供至少1个“虚拟IP”,作为接入IP。
本例中,如上表10.120.26.111。
其他
提前在2台服务器上部署好nginx(此处不演示)。
环境准备
下载官方安装包
如:
wget http://www.keepalived.org/software/keepalived-1.3.9.tar.gz
安装keepalive
本例使用root用户进行操作。
先执行如下指令,进行环境配置,并编译安装:
# 解压
tar xvf keepalived-1.3.9.tar.gz
cd keepalived-1.3.9
# 初始化环境
./configure --prefix=/home/keepalived
# 编译,并将可执行程序安装到/home/keepalived下
make && make install
--prefix,指定安装的路径,即可。
在执行configure过程中,可能出现如下错误:
1、缺少libnl等包,报错如下:
*** WARNING - this build will not support IPVS with IPv6. Please install libnl/libnl-3 dev libraries to support IPv6 with IPVS.
此时需要yum安装,如下:
# yum -y install libnl libnl-devel
2、缺少libnfnetlink包,报错:
configure: error: libnfnetlink headers missing
执行yum安装:
# yum install -y libnfnetlink-devel
防火墙添加vrrp组播规则
执行如下指令:
firewall-cmd --direct --permanent --add-rule ipv4 filter INPUT 0 --in-interface ens160 --destination 224.0.0.18 --protocol vrrp -j ACCEPT
firewall-cmd --reload
注意:
keepalive是基于vrrp协议工作的,因此必须配置组播规则。
关闭selinux
vi /etc/sysconfig/selinux
#修改:
SELINUX=disabled
#setenforce 0
说明:
开启selinux时,系统的安全性要求较高,可能导致keepvlive失效。
设置开启启动
拷贝文件:
cp /home/keepalived/etc/sysconfig/keepalived /etc/sysconfig/
cp /home/keepalived/sbin/keepalived /etc/init.d/
cp /home/keepalived/etc/keepalived/keepalived.conf /etc/keepalived/
说明:缺少的目录需要自行创建。
配置
keepvlive分为2种工作模式:抢占模式和非抢占模式,通过配置文件中的配置项state区分。
抢占模式
当Master宕机、恢复后,会抢夺Master角色,将自己升级为Master。
需要在配置文件中指定抢占模式,编辑/etc/keepalived/keepalived.conf配置文件:
1、MASTER配置(10.120.26.90)
global_defs {
router_id status #标识本节点的名称,通常为hostname
}
## keepalived会定时执行脚本并对脚本执行的结果进行分析,动态调整vrrp_instance的优先级。
##如果脚本执行结果为0,并且weight配置的值大于0,则优先级相应的增加。如果脚本执行结果非0,
##并且weight配置的值小于 0,则优先级相应的减少。其他情况,维持原本配置的优先级,即配置文件中priority对应的值。
vrrp_script chk_nginx {
script "/etc/keepalived/nginx_check.sh"
interval 2 #每2秒检测一次nginx的运行状态
weight -20 #失败一次,将自己的优先级-20
}
vrrp_instance VI_1 {
state MASTER # 状态,主节点为MASTER,备份节点为BACKUP
interface ens160 # 绑定VIP的网络接口,通过ifconfig查看自己的网络接口
virtual_router_id 51 # 虚拟路由的ID号,两个节点设置必须一样,可选IP最后一段使用,相同的VRID为一个组,他将决定多播的MAC地址
mcast_src_ip 10.120.26.90 # 本机IP地址
priority 100 # 节点优先级,值范围0~254,MASTER要比BACKUP高
advert_int 1 # 组播信息发送时间间隔,两个节点必须设置一样,默认为1秒
# 设置验证信息,两个节点必须一致
authentication {
auth_type PASS
auth_pass 1111
}
# 虚拟IP,两个节点设置必须一样。可以设置多个,一行写一个
virtual_ipaddress {
10.120.26.111
}
track_script {
chk_nginx # nginx存活状态检测脚本,对应vrrp_script配置块
}
}
2、 BACKUP配置(10.120.26.91)
global_defs {
router_id LB2 # 差异点1:BACKUP的主机名
}
vrrp_script chk_nginx {
script "/etc/keepalived/nginx_check.sh"
interval 2
weight -20
}
vrrp_instance VI_1 {
state BACKUP # 差异点2:路由器状态为BACKUP
interface ens160 # 根据本机的网卡名称填写
virtual_router_id 51
mcast_src_ip 10.120.26.91 # 差异点3:本机ip
priority 90 # 差异点4:要小于Master的优先级
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
10.120.26.111
}
track_script {
chk_nginx
}
}
非抢占模式
虚拟路由器组中的master从故障中恢复后,不会抢占备份节点的vip。
1、MASTER配置(10.120.26.90)
global_defs {
router_id LB1
}
vrrp_script chk_nginx {
script "/etc/keepalived/nginx_check.sh"
interval 2
weight -20
}
vrrp_instance VI_1 {
#state MASTER
state BACKUP # 注意这里的不同!
interface ens160
virtual_router_id 51
mcast_src_ip 10.120.26.90
priority 100 # 非抢占模式下通过priority判断是否为Master
advert_int 1
nopreempt
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
10.120.26.111
}
track_script {
chk_nginx
}
}
2、BACKUP(10.120.26.91)
除了优先级不一样,其余配置与Master一致。
nginx服务检测脚本
分别在主备服务器/etc/keepalived 目录下创建nginx_check.sh脚本,并为其添加执行权限:
/etc/keepalived
touch nginx_check.sh
chmod +x /etc/keepalived/nginx_check.sh
nginx_check.sh用于keepalived定时检测nginx的服务状态,如果nginx停止了,会尝试重新启动nginx,如果启动失败,会将keepalived进程杀死,将vip漂移到备份机器上。脚本内容如下:
#!/bin/bash
A=`ps -C nginx --no-header | wc -l`
if [ $A -eq 0 ];then
#尝试重新启动nginx
/opt/webapps/nginx/nginx_lr/sbin/nginx
sleep 2 #睡眠2秒
if [ `ps -C nginx --no-header | wc -l` -eq 0 ];then
systemctl stop keepalived.service #启动失败,将keepalived服务杀死。将vip漂移到其它备份节点
fi
fi
启动和验证
keepalive服务启动
systemctl start keepalived.service
查看进程状态:
[root@status etc]# ps -ef|grep keepalived
root 127479 1 0 17:58 ? 00:00:00 /home/keepalived/sbin/keepalived -D
root 127480 127479 0 17:58 ? 00:00:00 /home/keepalived/sbin/keepalived -D
root 127481 127479 0 17:58 ? 00:00:00 /home/keepalived/sbin/keepalived -D
root 127561 117202 0 17:58 pts/3 00:00:00 grep --color=auto keepalived
没问题后,分别查看两个节点的ip地址。Master节点:
[root@status etc]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000
link/ether 00:50:56:93:18:63 brd ff:ff:ff:ff:ff:ff
inet 10.120.26.90/24 brd 10.120.26.255 scope global ens160
valid_lft forever preferred_lft forever
inet 10.120.26.111/32 scope global ens160
valid_lft forever preferred_lft forever
inet6 fe80::250:56ff:fe93:1863/64 scope link
valid_lft forever preferred_lft forever
可以看到,10.120.26.90拿到了VIP 10.120.26.111。
BACKUP的ip情况:
[root@LB2 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000
link/ether 00:50:56:93:6e:eb brd ff:ff:ff:ff:ff:ff
inet 10.120.26.91/24 brd 10.120.26.255 scope global ens160
valid_lft forever preferred_lft forever
inet6 fe80::250:56ff:fe93:6eeb/64 scope link
valid_lft forever preferred_lft forever
BACKUP上只有10.120.26.91这个ip,正常。
验证
VIP漂移
我们先关闭Master,验证VIP是否会漂移到BACKUP上。
关闭 10.120.26.90 的 keepalived:
systemctl stop keepalived.service
查看 10.120.26.91 的ip状态:
[root@LB2 ~]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000
link/ether 00:50:56:93:6e:eb brd ff:ff:ff:ff:ff:ff
inet 10.120.26.91/24 brd 10.120.26.255 scope global ens160
valid_lft forever preferred_lft forever
inet 10.120.26.111/32 scope global ens160
valid_lft forever preferred_lft forever
inet6 fe80::250:56ff:fe93:6eeb/64 scope link
valid_lft forever preferred_lft forever
10.120.26.91 的 keepalived日志如下:
Dec 12 20:23:07 LB2 Keepalived_vrrp[75622]: VRRP_Instance(VI_1) Transition to MASTER STATE
Dec 12 20:23:08 LB2 Keepalived_vrrp[75622]: VRRP_Instance(VI_1) Entering MASTER STATE
Dec 12 20:23:08 LB2 Keepalived_vrrp[75622]: VRRP_Instance(VI_1) setting protocol VIPs.
Dec 12 20:23:08 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:08 LB2 Keepalived_vrrp[75622]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens160 for 10.120.26.111
Dec 12 20:23:08 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:08 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:08 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:08 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:13 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:13 LB2 Keepalived_vrrp[75622]: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on ens160 for 10.120.26.111
Dec 12 20:23:13 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:13 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:13 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
Dec 12 20:23:13 LB2 Keepalived_vrrp[75622]: Sending gratuitous ARP on ens160 for 10.120.26.111
看到“Transition to MASTER STATE”和“Entering MASTER STATE”,说明切换成功。
抢占模式下,VIP恢复
在抢占模式下,如果10.120.26.90(配置项 state=Master)恢复了,那么VIP就会漂移回来。
重启 10.120.26.90 的 keepalived之后,可以看到VIP又回来了:
ens160: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP qlen 1000
link/ether 00:50:56:93:18:63 brd ff:ff:ff:ff:ff:ff
inet 10.120.26.90/24 brd 10.120.26.255 scope global ens160
valid_lft forever preferred_lft forever
inet 10.120.26.111/32 scope global ens160
valid_lft forever preferred_lft forever
inet6 fe80::250:56ff:fe93:1863/64 scope link
valid_lft forever preferred_lft forever
这是,10.120.26.91 的 keepalived日志如下:
Dec 12 20:27:31 LB2 Keepalived_vrrp[75622]: VRRP_Instance(VI_1) Received advert with higher priority 100, ours 90
Dec 12 20:27:31 LB2 Keepalived_vrrp[75622]: VRRP_Instance(VI_1) Entering BACKUP STATE
Dec 12 20:27:31 LB2 Keepalived_vrrp[75622]: VRRP_Instance(VI_1) removing protocol VIPs.
到此,说明我们keepalive配置的没问题了。在Master路由器出现问题时,可以切换到BACKUP。
常见问题
1、主机访问不到nginx的80端口
报错:no route to host....
查看防火墙状态:
[root@LB1 test]# firewall-cmd --zone=public --list-ports
80/tcp
有这一行,表示80端口开放。没有的话需要添加规则:
firewall-cmd --zone=public --add-port=80/tcp --permanent
firewall-cmd --reload
2、抢占模式配置好后,两台机器都变成主机,都占有虚拟ip地址
这种情况可能是防火墙的组播配置没配置好,重新配置一下组播策略。