LVS+keepalived做Mysql集群的负载均衡:
注:
1、LVS本身不对node做健康检查的,所以要借助keepalived
2、根据配置当某node down掉会有几秒判断时间,这点请注意。
负载均衡端:
yum install ipvsadm keepalived ipset-devel
注:
keepalived的log在/var/log/message里
vi /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
notification_email {br/>aa@ming.com
}
notification_email_from aa@ming.com
smtp_server 10.0.0.10
smtp_connect_timeout 30
router_id LVS_DEVEL
vrrp_skip_check_adv_addr
vrrp_strict (这行一定要注掉,不注掉vip可能会ping不通)
vrrp_garp_interval 0
vrrp_gna_interval 0
}
vrrp_instance VI_1 {
state MASTER
interface eno1 (这里网卡名要写对)
virtual_router_id 51
priority 100
advert_int 3
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
10.0.1.10
}
}
virtual_server 10.0.1.10 3306 {
delay_loop 6
lb_algo rr
lb_kind DR (此处需改为DR)
persistence_atimeout 10
protocol TCP
real_server 10.0.1.2 3306 {
weight 1
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
connect_port 3306
}
}
real_server 10.0.1.3 3306 {
weight 1
TCP_CHECK {
connect_timeout 3
retry 3
delay_before_retry 1
connect_port 3306
}
}
real_server 10.0.1.4 3306 {
weight 1
TCP_CHECK {
connect_timeout 3
retry 3
delay_before_retry 1
connect_port 3306
}
}
}
:wq
systemctl restart keepalived
systemctl enable keepalived
systemctl status keepalived
ip a (检查eno1是不是多了个10.0.1.10的vip,这个ip在ifconfig是看不到的,要确保可以ping通)
ipvsadm -L --stats (查看各node分配情况)
Mysql node端(所有node都做):
ifconfig lo:0 10.0.1.10 netmask 255.255.255.255 broadcast 10.0.1.10
route add -host 10.0.1.10 dev lo:0
vi /etc/sysctl.conf
net.ipv4.conf.lo.arp_ignore = 1
net.ipv4.conf.lo.arp_announce = 2
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
:wq
sysctl -p
ifconfig lo:0
route -n
验证(可以看到访问的为不同node):
mysql -u aa -p'xxxx' -h 10.0.1.10 -e "select @@hostname;"
来源:oschina
链接:https://my.oschina.net/u/4299953/blog/4307501