不知道是不是架构师方面觉得服务器越少就越省钱,反正原先的4台服务器作一个服务的负载都觉得浪费。所以就出现了两台服务器作负载和服务的需求。
研究做过几次尝试
1.keepalived+ipvsadm+http,负载会失败。
2.keepalived+http,只能作vip,主备形式提供,流量只会到keepalive主的服务器上。只有障碍切换才会跑到备的主机上。
3.keepalived+nginx+http,这样可以提供轮询式的访问,但是要注意的是客户端无法再用http原来的端口访问,要在nginx上改变访问的端口。所以这种省钱方式其实还是不靠谱的。
但如果是前端程序访问后端服务可以自定义访问端口的情况下,这种方式起码让自己的成本帐面上省了一半
1.环境说明
ip | 作用 |
---|---|
192.168.20.99 | vip |
192.168.20.11 | keepalived(主)+nginx+http服务器 |
192.168.20.12 | keepalived(备)+nginx+http服务器 |
2.keepalived,nginx,http的安装
在有yum源的情况下,keepalived和httpd都可以通过yum安装
yum install -y keepalived httpd
nginx稍微麻烦一点,需要编译
编译环境安装
yum -y install make zlib zlib-devel gcc-c++ libtool openssl openssl-devel
下载nginx源码包并安装,这里要用到upstream这个模块,地址为http://nginx.org/en/download.html
tar -xzvf nginx-1.15.6.tar.gz
cd nginx-1.15.6
./configure --with-stream && make && make install
ln -s /usr/local/nginx/sbin/nginx /usr/bin/nginx
nginx -v #正确显示版本号以后说明nginx可以正常使用
keepalived的配置文件
192.168.20.11(主)上的配置文件
vip="192.168.20.99"
cat <<EOF>/etc/keepalived/keepalived.conf
vrrp_instance VI_1 {
state MASTER
interface enp0s8
virtual_router_id 99
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
$vip
}
}
EOF
192.168.20.12(备)上的配置文件
vip="192.168.20.99"
cat <<EOF>/etc/keepalived/keepalived.conf
vrrp_instance VI_1 {
state BACKUP
interface enp0s8
virtual_router_id 99
priority 80
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
$vip
}
}
EOF
启动keepalived并加入开机启动
systemctl start keepalived && systemctl enable keepalived
keepalived备机添加本地回路
添加是为了备机的nginx有相应的ip可以监听,不然没有vip监听,nginx无法启动
ip addr add 192.168.20.99/32 dev lo
#可以加入开机启动
chmod 755 /etc/rc.d/rc.local
echo "ip addr add 192.168.20.99 dev lo">>/etc/rc.local
httpd的主页文件(为了方便测试,把主机名写入index.html)
echo "my name is `hostname`">/var/www/html/index.html
启动httpd
systemctl start httpd && systemctl enable httpd
nginx的配置文件
以下配置文件中前端访问端口变为了88,如果还是80的话,因为http 80 占用,nginx是无法绑定的。
cat <<EOF>/usr/local/nginx/conf/nginx.conf
worker_processes 1;
worker_rlimit_nofile 200000;
events {
worker_connections 120000;
}
stream {
include /etc/nginx/stream/*.stream;
}
EOF
[ -d /etc/nginx/stream ] || mkdir -p /etc/nginx/stream
cat <<EOF>/etc/nginx/stream/192.168.20.99_80.stream
upstream 192.168.20.99_80 {
server 192.168.20.11:80 weight=10;
server 192.168.20.12:80 weight=10;
}
server {
listen 192.168.20.99:88;
proxy_connect_timeout 1s;
proxy_timeout 30s;
proxy_pass 192.168.20.99_80;
}
EOF
语法检查,如果没有问题就启动nginx
nginx -t
nginx
# 加入开机启动项
chmod 755 /etc/rc.d/rc.local
echo "nginx" >>/etc/rc.local
3.测试
注意访问的是88端口
myname@mypc:~/testbox$ for i in `seq 1 5`;do curl 192.168.20.99:88;done
my name is test1
my name is test2
my name is test1
my name is test2
my name is test1
网友评论