Keepalived 软件起初是专为LVS负载均衡软件设计的,用来管理并监控LVS集群系统中各个服务节点的状态,后来又加入了可以实现高可用的VRRP功能。因此,Keepalived除了能够管理LVS软件外,还可以作为其他服务(例如:Nginx、Haproxy、MySQL等)的高可用解决方案软件。
Keepalived软件主要是通过VRRP协议实现高可用功能的。VRRP是Virtual Router RedundancyProtocol(虚拟路由器冗余协议)的缩写,VRRP出现的目的就是为了解决静态路由单点故障问题的,它能够保证当个别节点宕机时,整个网络可以不间断地运行。
所以,Keepalived 一方面具有配置管理LVS的功能,同时还具有对LVS下面节点进行健康检查的功能,另一方面也可实现系统网络服务的高可用功能。
keepalived 有三个重要的功能,分别是:
Keepalived 高可用服务之间的故障切换转移,是通过 VRRP (Virtual Router Redundancy Protocol ,虚拟路由器冗余协议)来实现的。
在 Keepalived 服务正常工作时,主 Master 节点会不断地向备节点发送(多播的方式)心跳消息,用以告诉备 Backup 节点自己还活看,当主 Master 节点发生故障时,就无法发送心跳消息,备节点也就因此无法继续检测到来自主 Master 节点的心跳了,于是调用自身的接管程序,接管主 Master 节点的 IP 资源及服务。而当主 Master 节点恢复时,备 Backup 节点又会释放主节点故障时自身接管的IP资源及服务,恢复到原来的备用角色。
那么,什么是VRRP呢?
VRRP ,全 称 Virtual Router Redundancy Protocol ,中文名为虚拟路由冗余协议 ,VRRP的出现就是为了解决静态踣甶的单点故障问题,VRRP是通过一种竞选机制来将路由的任务交给某台VRRP路由器的。
Keepalived高可用对之间是通过VRRP通信的,因此,我们从 VRRP开始了解起:
介绍完 VRRP,接下来我再介绍一下 Keepalived服务的工作原理:
Keepalived高可用是通过 VRRP 进行通信的, VRRP是通过竞选机制来确定主备的,主的优先级高于备,因此,工作时主会优先获得所有的资源,备节点处于等待状态,当主挂了的时候,备节点就会接管主节点的资源,然后顶替主节点对外提供服务。
在 Keepalived 服务之间,只有作为主的服务器会一直发送 VRRP 广播包,告诉备它还活着,此时备不会枪占主,当主不可用时,即备监听不到主发送的广播包时,就会启动相关服务接管资源,保证业务的连续性.接管速度最快可以小于1秒。
keepalived 的主配置文件是 /etc/keepalived/keepalived.conf。其内容如下:
root@master ~]# cat /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs { //全局配置
notification_email { //定义报警收件人邮件地址
[email protected]
[email protected]
[email protected]
}
notification_email_from [email protected] //定义报警发件人邮箱
smtp_server 192.168.200.1 //邮箱服务器地址
smtp_connect_timeout 30 //定义邮箱超时时间
router_id LVS_DEVEL //定义路由标识信息,同局域网内唯一
vrrp_skip_check_adv_addr
vrrp_strict
vrrp_garp_interval 0
vrrp_gna_interval 0
}
vrrp_instance VI_1 { //定义实例
state MASTER //指定keepalived节点的初始状态,可选值为MASTER|BACKUP
interface eth0 //VRRP实例绑定的网卡接口,用户发送VRRP包
virtual_router_id 51 //虚拟路由的ID,同一集群要一致
priority 100 //定义优先级,按优先级来决定主备角色,优先级越大越优先
nopreempt //设置不抢占
advert_int 1 //主备通讯时间间隔
authentication { //配置认证
auth_type PASS //认证方式,此处为密码
auth_pass 1111 //同一集群中的keepalived配置里的此处必须一致,推荐使用8位随机数
}
virtual_ipaddress { //配置要使用的VIP地址
192.168.200.16
}
}
virtual_server 10.10.10.2 1358 { //配置虚拟服务器
delay_loop 6 //健康检查的时间间隔
lb_algo rr //lvs调度算法
lb_kind NAT //lvs模式
persistence_timeout 50 //持久化超时时间,单位是秒
protocol TCP //4层协议
sorry_server 192.168.200.200 1358 //定义备用服务器,当所有RS都故障时用sorry_server来响应客户端
real_server 192.168.200.2 1358 { //定义真实处理请求的服务器
weight 1 //给服务器指定权重,默认为1
HTTP_GET {
url {
path /testurl/test.jsp //指定要检查的URL路径
digest 640205b7b0fc66c1ea91c463fac6334d //摘要信息
}
url {
path /testurl2/test.jsp
digest 640205b7b0fc66c1ea91c463fac6334d
}
url {
path /testurl3/test.jsp
digest 640205b7b0fc66c1ea91c463fac6334d
}
connect_timeout 3 //连接超时时间
nb_get_retry 3 //get尝试次数
delay_before_retry 3 //在尝试之前延迟多长时间
}
}
real_server 192.168.200.3 1358 {
weight 1
HTTP_GET {
url {
path /testurl/test.jsp
digest 640205b7b0fc66c1ea91c463fac6334c
}
url {
path /testurl2/test.jsp
digest 640205b7b0fc66c1ea91c463fac6334c
}
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}
vrrp_instance段配置
nopreempt //设置为不抢占。默认是抢占的,当高优先级的机器恢复后,会抢占低优先 \
级的机器成为MASTER,而不抢占,则允许低优先级的机器继续成为MASTER,即使高优先级 \
的机器已经上线。如果要使用这个功能,则初始化状态必须为BACKUP。
preempt_delay //设置抢占延迟。单位是秒,范围是0---1000,默认是0.发现低优先 \
级的MASTER后多少秒开始抢占。
vrrp_script段配置
weight //给服务器指定权重。默认是1
inhibit_on_failure //当服务器健康检查失败时,将其weight设置为0, \
而不是从Virtual Server中移除
notify_up //当服务器健康检查成功时,执行的脚本
notify_down //当服务器健康检查失败时,执行的脚本
uthreshold //到这台服务器的最大连接数
lthreshold //到这台服务器的最小连接数
tcp_check段配置
connect_ip //连接的IP地址。默认是real server的ip地址
connect_port //连接的端口。默认是real server的端口
bindto //发起连接的接口的地址。
bind_port //发起连接的源端口。
connect_timeout //连接超时时间。默认是5s。
fwmark //使用fwmark对所有出去的检查数据包进行标记。
warmup //指定一个随机延迟,最大为N秒。可防止网络阻塞。如果为0,则关闭该功能。
retry //重试次数。默认是1次。
delay_before_retry //默认是1秒。在重试之前延迟多少秒。
global_defs {
router_id LVS_Server
}
vrrp_instance VI_1 {
state BACKUP
interface ens33
virtual_router_id 51
priority 150
nopreempt
advert_int 1
authentication {
auth_type PASS
auth_pass wangqing
}
virtual_ipaddress {
172.16.12.250 dev ens33
}
}
virtual_server 172.16.12.250 80 {
delay_loop 3
lvs_sched rr
lvs_method DR
protocol TCP
real_server 172.16.12.129 80 {
weight 1
TCP_CHECK {
connect_port 80
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 172.16.12.130 8080 {
weight 1
TCP_CHECK {
connect_port 8080
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}
在高可用(HA)系统中,当联系2个节点的“心跳线”断开时,本来为一整体、动作协调的HA系统,就分裂成为2个独立的个体。由于相互失去了联系,都以为是对方出了故障。两个节点上的HA软件像“裂脑人”一样,争抢“共享资源”、争起“应用服务”,就会发生严重后果——或者共享资源被瓜分、2边“服务”都起不来了;或者2边“服务”都起来了,但同时读写“共享存储”,导致数据损坏(常见如数据库轮询着的联机日志出错)。
对付HA系统“裂脑”的对策,目前达成共识的的大概有以下几条:
一般来说,脑裂的发生,有以下几种原因:
注意:
Keepalived配置里同一 VRRP实例如果 virtual_router_id两端参数配置不一致也会导致裂脑问题发生。
在实际生产环境中,我们可以从以下几个方面来防止裂脑问题的发生:
当然,在实施高可用方案时,要根据业务实际需求确定是否能容忍这样的损失。对于一般的网站常规业务.这个损失是可容忍的
对脑裂的监控应在备用服务器上进行,通过添加zabbix自定义监控进行。
监控什么信息呢?监控备上有无VIP地址
备机上出现VIP有两种情况:
监控只是监控发生脑裂的可能性,不能保证一定是发生了脑裂,因为正常的主备切换VIP也是会到备上的。
监控脚本如下:
[root@slave ~]# mkdir -p /scripts && cd /scripts
[root@slave scripts]# vim check_keepalived.sh
#!/bin/bash
while true
do
if [ `ip a show ens33 |grep 172.16.12.250|wc -l` -ne 0 ]
then
echo "keepalived is error!"
else
echo "keepalived is OK !"
fi
done
编写脚本时要注意,网卡要改成你自己的网卡名称,VIP也要改成你自己的VIP,最后不要忘了给脚本赋予执行权限,且要修改/scripts目录的属主属组为zabbix
环境说明
系统信息 | 主机名 | IP |
---|---|---|
rhel7.4 | master | 192.168.155.12 |
rhel7.4 | slave | 192.168.155.13 |
本次高可用虚拟IP(VIP)地址暂定为 192.168.155.200
配置主keepalived
//关闭防火墙与SELINUX
[root@master ~]# systemctl stop firewalld
[root@master ~]# systemctl disable firewalld
Removed symlink /etc/systemd/system/multi-user.target.wants/firewalld.service.
Removed symlink /etc/systemd/system/dbus-org.fedoraproject.FirewallD1.service.
[root@master ~]# setenforce 0
[root@master ~]# sed -ri 's/^(SELINUX=).*/\1disabled/g' /etc/selinux/config
//配置网络源
[root@master ~]# curl -o /etc/yum.repos.d/CentOS7-Base-163.repo http://mirrors.163.com/.help/CentOS7-Base-163.repo
[root@master ~]# sed -i 's/\$releasever/7/g' /etc/yum.repos.d/CentOS7-Base-163.repo
[root@master ~]# sed -i 's/^enabled=.*/enabled=1/g' /etc/yum.repos.d/CentOS7-Base-163.repo
[root@master ~]# yum -y install epel-release vim wget gcc gcc-c++
安装过程略.....
//安装keepalived
[root@master ~]# yum -y install keepalived
//查看安装生成的文件
[root@master ~]# rpm -ql keepalived
/etc/keepalived //配置目录
/etc/keepalived/keepalived.conf //此为主配置文件
/etc/sysconfig/keepalived
/usr/bin/genhash
/usr/lib/systemd/system/keepalived.service //此为服务控制文件
/usr/libexec/keepalived
/usr/sbin/keepalived
.....此处省略N行
用同样的方法在备服务器上安装keepalived
//关闭防火墙与SELINUX
[root@slave ~]# systemctl stop firewalld
[root@slave ~]# systemctl disable firewalld
Removed symlink /etc/systemd/system/multi-user.target.wants/firewalld.service.
Removed symlink /etc/systemd/system/dbus-org.fedoraproject.FirewallD1.service.
[root@slave ~]# setenforce 0
[root@slave ~]# sed -ri 's/^(SELINUX=).*/\1disabled/g' /etc/selinux/config
//配置网络源
[root@slave ~]# curl -o /etc/yum.repos.d/CentOS7-Base-163.repo http://mirrors.163.com/.help/CentOS7-Base-163.repo
[root@slave ~]# sed -i 's/\$releasever/7/g' /etc/yum.repos.d/CentOS7-Base-163.repo
[root@slave ~]# sed -i 's/^enabled=.*/enabled=1/g' /etc/yum.repos.d/CentOS7-Base-163.repo
[root@slave ~]# yum -y install epel-release vim wget gcc gcc-c++
安装过程略.....
//安装keepalived
[root@slave ~]# yum -y install keepalived
在做之前需关闭备keepalived上的mysql服务
[root@master ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id lb01
}
vrrp_instance VI_1 {
state MASTER
interface ens32 //自己本机网卡
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 11111111 //与slave上密码要一致
}
virtual_ipaddress {
192.168.155.200 //虚拟ip
}
}
virtual_server 192.168.155.200 {
delay_loop 6
lb_algo rr
lb_kind DR
persistence_timeout 50
protocol TCP
real_server 192.168.155.12 3306 { //mysql服务端口,下面也是一样
weight 1
TCP_CHECK {
connect_port 3306
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.155.13 3306 {
weight 1
TCP_CHECK {
connect_port 3306
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}
[root@master ~]# systemctl start keepalived
[root@master ~]# systemctl enable keepalived
[root@slave ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id lb02
}
vrrp_instance VI_1 {
state BACKUP
interface ens32 //自己本机网卡
virtual_router_id 51
priority 90
advert_int 1
authentication {
auth_type PASS
auth_pass 11111111 //与master上密码要一致
}
virtual_ipaddress {
192.168.155.200 //虚拟ip,与master上一致
}
}
virtual_server 192.168.155.200 3306 {
delay_loop 6
lb_algo rr
lb_kind DR
persistence_timeout 50
protocol TCP
real_server 192.168.155.12 3306 { //master的ip
weight 1
TCP_CHECK {
connect_port 3306
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.155.13 3306 { //slave的ip
weight 1
TCP_CHECK {
connect_port 3306
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}
[root@slave ~]# systemctl start keepalived
[root@slave ~]# systemctl enable keepalived
在MASTER上查看
[root@master ~]# ip a
1: lo: 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: ens32: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:a5:fd:35 brd ff:ff:ff:ff:ff:ff
inet 192.168.155.12/24 brd 192.168.155.255 scope global ens32
valid_lft forever preferred_lft forever
inet 192.168.155.200/32 scope global ens32 //可以看到此处有VIP
valid_lft forever preferred_lft forever
inet6 fe80::b689:664:a986:e2f7/64 scope link
valid_lft forever preferred_lft forever
在SLAVE上查看
[root@slave ~]# ip a
1: lo: 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: ens32: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:5d:50:53 brd ff:ff:ff:ff:ff:ff
inet 192.168.155.13/24 brd 192.168.155.255 scope global ens32
valid_lft forever preferred_lft forever
inet6 fe80::b2fa:92cf:a311:1389/64 scope link
valid_lft forever preferred_lft forever
此步可做可不做,该功能可用于仅监听VIP的时候
在master上修改内核参数
[root@master ~]# echo 'net.ipv4.ip_nonlocal_bind = 1' >>/etc/sysctl.conf
[root@master ~]# sysctl -p
net.ipv4.ip_nonlocal_bind = 1
[root@master ~]# cat /proc/sys/net/ipv4/ip_nonlocal_bind
1
在slave上修改内核参数
[root@slave ~]# echo 'net.ipv4.ip_nonlocal_bind = 1' >>/etc/sysctl.conf
[root@slave ~]# sysctl -p
net.ipv4.ip_nonlocal_bind = 1
[root@slave ~]# cat /proc/sys/net/ipv4/ip_nonlocal_bind
1
keepalived通过脚本来监控mysql状态
在master上编写脚本
[root@master ~]# mkdir /scripts
[root@master ~]# cd /scripts/
[root@master scripts]# vim check_m.sh
#!/bin/bash
mysql_status=$(ps -ef|grep -Ev "grep|$0"|grep '\bmysql\b'|wc -l)
if [ $mysql_status -lt 1 ];then
systemctl stop keepalived
fi
[root@master scripts]# chmod +x check_m.sh
[root@master scripts]# ll
total 4
-rwxr-xr-x 1 root root 168 Oct 19 23:38 check_m.sh
[root@master scripts]# vim notify.sh
#!/bin/bash
VIP=$2
sendmail (){
subject="${VIP}'s server keepalived state is translate"
content="`date +'%F %T'`: `hostname`'s state change to master"
echo $content | mail -s "$subject" [email protected]
}
case "$1" in
master)
mysql_status=$(ps -ef|grep -Ev "grep|$0"|grep '\bmysql\b'|wc -l)
if [ $mysql_status -lt 1 ];then
systemctl start mysqld
fi
sendmail
;;
backup)
mysql_status=$(ps -ef|grep -Ev "grep|$0"|grep '\bmysql\b'|wc -l)
if [ $nginx_status -gt 0 ];then
systemctl stop mysqld
fi
;;
*)
echo "Usage:$0 master|backup VIP"
;;
esac
[root@master scripts]# chmod +x notify.sh
[root@master scripts]# ll
total 8
-rwxr-xr-x 1 root root 168 Oct 19 23:38 check_n.sh
-rwxr-xr-x 1 root root 594 Oct 20 03:24 notify.sh
在slave上编写脚本
[root@slave ~]# mkdir /scripts
[root@slave ~]# cd /scripts/
[root@slave scripts]# vim notify.sh
#!/bin/bash
VIP=$2
sendmail (){
subject="${VIP}'s server keepalived state is translate"
content="`date +'%F %T'`: `hostname`'s state change to master"
echo $content | mail -s "$subject" [email protected]
}
case "$1" in
master)
mysql_status=$(ps -ef|grep -Ev "grep|$0"|grep '\bmysql\b'|wc -l)
if [ $mysql_status -lt 1 ];then
systemctl start mysqld
fi
sendmail
;;
backup)
mysql_status=$(ps -ef|grep -Ev "grep|$0"|grep '\bmysql\b'|wc -l)
if [ $mysql_status -gt 0 ];then
systemctl stop mysqld
fi
;;
*)
echo "Usage:$0 master|backup VIP"
;;
esac
[root@slave scripts]# chmod +x notify.sh
[root@slave scripts]# ll
total 8
-rwxr-xr-x 1 root root 168 Oct 19 23:38 check_n.sh
-rwxr-xr-x 1 root root 594 Oct 20 03:24 notify.sh
此处的脚本名称应避免与服务名相同,推荐用服务名的首字母代替,如check_m,不要给脚本起名check_mysqld
配置主keepalived
[root@master ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id lb01
}
vrrp_script mysql_check {
script "/scripts/check_m.sh"
interval 1
weight -20
}
vrrp_instance VI_1 {
state MASTER
interface ens32
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 11111111
}
virtual_ipaddress {
192.168.155.200
}
track_script {
mysql_check
}
notify_master "/scripts/notify.sh master 192.168.155.200"
notify_backup "/scripts/notify.sh backup 192.168.155.200"
}
virtual_server 192.168.155.200 3306 {
delay_loop 6
lb_algo rr
lb_kind DR
persistence_timeout 50
protocol TCP
real_server 192.168.155.12 3306 {
weight 1
TCP_CHECK {
connect_port 3306
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 172.16.12.13 3306 {
weight 1
TCP_CHECK {
connect_port 3306
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}
[root@master ~]# systemctl restart keepalived
配置备keepalived
backup无需检测nginx是否正常,当升级为MASTER时启动nginx,当降级为BACKUP时关闭
[root@slave ~]# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id lb02
}
vrrp_instance VI_1 {
state BACKUP
interface ens32
virtual_router_id 51
priority 90
nopreempt
advert_int 1
authentication {
auth_type PASS
auth_pass 11111111
}
virtual_ipaddress {
192.168.155.200
}
notify_master "/scripts/notify.sh master 192.168.155.200"
notify_backup "/scripts/notify.sh backup 192.168.155.200"
}
virtual_server 192.168.155.200 3306 {
delay_loop 6
lb_algo rr
lb_kind DR
persistence_timeout 50
protocol TCP
real_server 192.168.155.12 3306 {
weight 1
TCP_CHECK {
connect_port 3306
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
real_server 192.168.155.12 3306 {
weight 1
TCP_CHECK {
connect_port 3306
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}
[root@slave ~]# systemctl restart keepalived
手动关闭主keepalived上的mysql服务,并执行脚本check_m.sh
[root@master scripts]# systemctl stop mysqld
[root@master scripts]# bash -x check_m.sh
++ ss -antl
++ grep 3306
++ wc -l
+ mysql_status=0
+ '[' 0 -lt 1 ']'
+ systemctl stop keepalived
[root@master scripts]# ip a
1: lo: 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: ens32: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:a5:fd:35 brd ff:ff:ff:ff:ff:ff
inet 192.168.155.12/24 brd 192.168.155.255 scope global ens32
valid_lft forever preferred_lft forever
inet6 fe80::b689:664:a986:e2f7/64 scope link
valid_lft forever preferred_lft forever
去备keepalived上查看,keepalive配置成功
[root@slave scripts]# ss -antl
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 128 *:22 *:*
LISTEN 0 100 127.0.0.1:25 *:*
LISTEN 0 128 :::22 :::*
LISTEN 0 100 ::1:25 :::*
[root@slave scripts]# ss -antl
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 128 *:22 *:*
LISTEN 0 100 127.0.0.1:25 *:*
LISTEN 0 128 :::22 :::*
LISTEN 0 100 ::1:25 :::*
LISTEN 0 80 :::3306 :::*
[root@slave scripts]# ip a
1: lo: 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: ens32: mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:0c:29:5d:50:53 brd ff:ff:ff:ff:ff:ff
inet 192.168.155.13/24 brd 192.168.155.255 scope global ens32
valid_lft forever preferred_lft forever
inet 192.168.155.200/32 scope global ens32 //主keepalived宕机之后,此备keepalived变成了主
valid_lft forever preferred_lft forever
inet6 fe80::b2fa:92cf:a311:1389/64 scope link
valid_lft forever preferred_lft forever