文章目录
keepalived 高可用
keepalived高可用只是用来企业内部，解决单点故障的软件
什么是高可用
一般是指2台机器启动着完全相同的业务系统，当有一台机器down机了，另外一台服务器就能快速的接管，对于访问的用户是无感知的。
keepalived是如何实现高可用的
keepalived软件是基于VRRP协议实现的，VRRP是虚拟路由冗余协议，主要用于解决单点故障问题
如何才能做到出现故障自动转移，此时VRRP就出现了，我们的VRRP其实是通过软件或者硬件的形式在Master和Backup外面增加一个虚拟的MAC地址（VMAC）与虚拟IP地址（VIP），那么在这种情况下，PC请求VIP的时候，无论是Master处理还是Backup处理，PC仅会在ARP缓存表中记录VMAC与VIP的信息。
高可用keepalived核心概念
1、如何确定谁是主节点谁是备节点（选举投票，优先级）
2、如果Master故障，Backup自动接管，那么Master恢复后会夺权吗（抢占试、非抢占式）
3、如果两台服务器都认为自己是Master会出现什么问题（脑裂）
keepalived 高可用安装配置
lb01负载均衡实现
#web端
[root@web01 conf.d]# vim linux.com.conf
server {
listen 80;
server_name _;
charset utf8;
location / {
root /code/node;
index index.html;
}
}
#web02 web03也建立目录并授权
[root@web01 conf.d]# mkdir /code/node -p
[root@web01 conf.d]# chown -R www.www /code
[root@web01 conf.d]# echo "我是web01......." > /code/node/index.html
[root@web02 conf.d]# echo "我是web02......." > /code/node/index.html
[root@web03 conf.d]# echo "我是web03....." > /code/node/index.html
#重启
[root@web01 conf.d]# !sy
systemctl restart nginx
#负载均衡lb01
[root@lb01 conf.d]# vim lb.conf
upstream http {
server 172.16.1.7:80;
server 172.16.1.8:80;
server 172.16.1.9:80;
}
server {
listen 443 ssl;
server_name _;
ssl_certificate /etc/nginx/ssl_key/server.crt;
ssl_certificate_key /etc/nginx/ssl_key/server.key;
location / {
proxy_pass http://http;
}
}
server {
listen 80;
server_name linux.lb.com;
rewrite (.*) https://$server_name$request_uri;
}
#加入hosts 并访问
环境准备
| 主机 | IP | 身份 |
|---|---|---|
| lb01 | 172.16.1.5 | master |
| lb02 | 172.16.1.6 | backup |
| 192.168.15.3 | VIP |
保证lb01和lb02配置完全一致
# lb01和lb02上的nginx配置挂载到nfs上
# 创建一个挂载点
[root@nfs nfs]# mkdir lb
# 授权
[root@nfs nfs]# chown www.www lb/
# 增加挂载点
[root@nfs nfs]# vim /etc/exports
/nfs/lb 172.16.1.0/20(rw,sync,all_squash,anonuid=1000,anongid=1000)
# 重启nfs服务
[root@nfs nfs]# systemctl restart nfs-server rpcbind
# 在lb01和lb02上挂载
[root@lb01 ~]# mount -t nfs 172.16.1.31:/nfs/lb /etc/nginx/conf.d/
# 编写配置
[root@lb02 ~]# cat /etc/nginx/conf.d/http.conf
upstream http {
server 172.16.1.7:80;
server 172.16.1.8:80;
server 172.16.1.9:80;
}
server {
listen 443 ssl;
server_name _;
ssl_certificate /etc/nginx/cert/server.crt;
ssl_certificate_key /etc/nginx/cert/server.key;
location / {
proxy_pass http://http;
}
}
server {
listen 80;
server_name 192.168.15.5;
rewrite (.*) https://$server_name$request_uri;
}
[root@lb02 ~]# systemctl restart nginx
安装keepalived
[root@lb01 ~]# yum install -y keepalived
[root@lb02 ~]# yum install -y keepalived
抢占式
配置keepalived主节点
#查看配置文件
[root@lb01 ~]# rpm -qc keepalived
/etc/keepalived/keepalived.conf
/etc/sysconfig/keepalived
#配置主节点配置文件
[root@lb01 ~]# vim /etc/keepalived/keepalived.conf
global_defs { #全局配置
router_id lb01 #身份验证
}
vrrp_instance VI_1 {
state MASTER #状态，只有MASTER和BACKUP，MASTER是主，BACKUP是备
interface eth0 #网卡绑定，心跳检测
virtual_router_id 51 #虚拟路由标识，组id，把master和backup判断为一组
priority 100 #优先级（真正判断是主是从的条件）（值越大优先级越高）
advert_int 3 #检测状态间隔时间（单位是秒）
authentication { #认证
auth_type PASS #认证方式
auth_pass 1111 #认证密码指定
}
virtual_ipaddress {
192.168.15.3 #虚拟的VIP地址
}
}
配置keepalived从节点
[root@lb02 ~]# vim /etc/keepalived/keepalived.conf
global_defs {
router_id lb02
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
virtual_router_id 51
priority 90
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.15.3
}
}
配置区别
| keepalived配置区别 | MASTER节点 | BACKUP节点 |
|---|---|---|
| router_id（唯一标识） | lb01 | lb02 |
| state（角色状态） | MASTER | BACKUP |
| priority（优先级） | 100 | 90 |
启动keepalived并查看
#先启动备节点
[root@lb02 ssl_key]# systemctl start keepalived.service
#查看ip
[root@lb02 ssl_key]# ip addr | grep 192.168.15.3
inet 192.168.15.3/32 scope global eth0
#启动主节点
[root@lb01 ssl_key]# systemctl start keepalived.service
#查看lb01的ip
[root@lb01 ssl_key]# ip addr | grep 192.168.15.3
inet 192.168.15.3/32 scope global eth0
#查看lb02的ip已经没有了192.168.15.3/32
keepalived绑定日志
#配置keepalived
[root@lb01 ~]# vim /etc/sysconfig/keepalived
KEEPALIVED_OPTIONS="-D -d -S 0" #其中-S指定syslog的facility
#配置rsyslog抓取日志
[root@lb01 ~]# vim /etc/rsyslog.conf
local0.* /var/log/keepalived.log
#重启服务
[root@lb01 ~]# systemctl restart keepalived
[root@lb01 ~]# systemctl restart rsyslog
keepalived是抢占式的创建虚拟VIP，抢占式创建虚拟VIP的时候，容易导致网络不稳定
非抢占式
我们一般配置的都是非抢占式的，因为宕机这种行为一次就够了
1.修改节点状态，两边状态都必须是BACKUP
2.两个节点都要加上 nopreempt
3.优先级仍保持不同 priority
#主节点配置
[root@lb01 ~]# cat /etc/keepalived/keepalived.conf
global_defs {
router_id lb01
}
vrrp_instance VI_1 {
state BACKUP # 如果配置非抢占式VIP，集群状态必须一致
interface eth0
virtual_router_id 51
priority 100
nopreempt # 配置非抢占式VIP
advert_int 3
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.15.3
#从节点配置
[root@lb02 ~]# cat /etc/keepalived/keepalived.conf
global_defs {
router_id lb02
}
vrrp_instance VI_1 {
state BACKUP # 如果配置非抢占式VIP，集群状态必须一致
interface eth0
virtual_router_id 51
priority 50
nopreempt # 配置非抢占式VIP
advert_int 3
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.15.3
}
}
高可用keepalived的脑裂
由于某些原因，导致两台keepalived高可用服务器在指定时间内，无法检测到对方是否存活，各自去调用资源，分配工作，而此时两台服务器都还活着并且在工作。
备用节点如何知道主节点是否down机？
备用节点一直在PING挂载在主节点的VIP.主节点的VIP会给备用节点会回复PONG，证明主节点并没有宕机。如果主节点没有回复，则备用节点则会启动自己的VIP。
脑裂的故障
1.服务器网线松动，网络故障
2.服务器硬件发生损坏，硬件故障
3.主备服务器之间开启了防火墙
开启防火墙
[root@lb01 ~]# systemctl start firewalld
[root@lb02 ~]# systemctl start firewalld
#访问浏览器因为开启防火墙，所以访问不了站点，需要配置开启http服务
[root@lb02 ~]# firewall-cmd --add-service=http
[root@lb02 ~]# firewall-cmd --add-service=https
解决keepalived的脑裂问题的办法
-eq 等于
-ne 不等于
-ge 大于等于
-gt 大于
-le 小于等于
-lt 小于
#干掉一台服务
[root@lb02 ~]# systemctl stop keepalived
#判断是否有脑裂现象
先做免密
[root@lb01 ~]# ssh-kengen -t rsa
[root@lb02 ~]# ssh-copy-id -i ~/.ssh/id_rsa.pub root@172.16.1.6
[root@lb02 ~]# ssh-kengen -t rsa
[root@lb02 ~]# ssh-copy-id -i ~/.ssh/id_rsa.pub root@172.16.1.5
# 当主节点和从节点都提供服务的时候（脚本探测）
[root@lb01 ~]# systemctl start firewalld
[root@lb02 ~]# systemctl start firewalld
#开启之后会出现脑裂现象
#在lb01部署
[root@lb02 ~]# cat cheak_vrrp.sh
#!/bin/bash
VIP="192.168.15.3"
MASTERIP="172.16.1.6"
BACKUPIP="172.16.1.5"
while true; do
PROBE='ip a | grep "${VIP}"'
ssh ${MASTERIP} "${PROBE}" > /dev/null
MASTER_STATU=$?
ssh ${BACKUPIP} "${PROBE}" > /dev/null
BACKUP_STATU=$?
if [[ $MASTER_STATU -eq 0 && $BACKUP_STATU -eq 0 ]];then
ssh ${BACKUPIP} "systemctl stop keepalived.service"
fi
sleep 2
done
nginx故障切换脚本
域名解析到VIP
nginx默认监听所有IP
两台机部署
#如果nginx宕机，用户请求页面会失败，但是keepalive没有关闭，VIP仍然在nginx挂掉了的机器上，导致影响业务；
#我们应该编写一个脚本，判断nginx状态，如果nginx挂掉，先尝试重启nginx，如果启动不了则关掉keepalived
# nginx检测脚本两台机部署
[root@lb01 ~]# cat web_check.sh
#!/bin/bash
nginxnum=`ps -ef | grep [n]ginx | wc -l`
if [ $nginxnum -eq 0 ];then
systemctl start nginx
sleep 3
nginxnum=`ps -ef | grep [n]ginx | wc -l`
if [ $nginxnum -eq 0 ];then
systemctl stop keepalived.service
fi
fi
#给脚本添加执行权限
[root@lb01 ~]# chmod +x web_check.sh
#停掉nginx执行脚本测试
[root@lb01 ~]# systemctl stop nginx
[root@lb01 ~]# ./web_check.sh
[root@lb01 ~]# systemctl restart keepalived.service
[root@1b02 ~]# systemctl restart keepalived.service
#结论脚本检测nginx是否坏掉 重启 重启不成功干掉keepalived 实现vip漂移无感知
#调用脚本
root@lb01 ~]# vim /etc/keepalived/keepalived.conf
global_defs {
router_id lb01
}
#每5秒执行一次脚本，脚本执行完成时间不能超过5秒，否则会重新执行脚本，死循环
vrrp_script check_web {
script "/root/web_check.sh"
interval 5
}
vrrp_instance VI_1 {
state MASTER
interface eth0
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.15.3
}
#调用计划脚本
track_script {
check_web
}
}
[root@1b02 ~]# cat /etc/keepalived/keepalived.conf
global_defs {
router_id lb02
}
vrrp_script check_web {
script "/root/web_check.sh"
interval 5
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
virtual_router_id 51
priority 50
nopreempt
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.15.3
}
track_script {
check_web
}
}
#给脚本添加执行权限执行（可以放在后台执行加上&）
[root@lb01 ~]# chmod +x check_web.sh
======================================================================
负载均衡----discuz
web端（01.02.03）
#挂载
[root@web01 conf.d]# mount -t nfs 172.16.1.31:/nfs/web /www/
[root@web01 conf.d]# mount -t nfs 172.16.1.31:/nfs/conf /etc/nginx/conf.d/
#创建证书并推送
[root@web01 ~]# mkdir /etc/nginx/ssl_key
[root@web01 ~]# cd /etc/nginx/ssl_key/
[root@web01 ssl_key]# openssl genrsa -idea -out server.key 2048
[root@web01 ssl_key]# openssl req -days 36500 -x509 -sha256 -nodes -newkey rsa:2048 -keyout server.key -out server.crt
[root@web01 conf.d]#scp -r /etc/nginx/ssl_key/ 172.16.1.5:/etc/nginx/
注：web02 web03 负载均衡都要推送
#统一用户 配置web01
[root@web01 conf.d]# vim discuz.conf
server {
listen 80;
server_name linux.discuz.com;
location / {
root /www/upload;
index index.php;
}
location ~* \.php$ {
root /www/upload;
fastcgi_pass 127.0.0.1:9000;
fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
fastcgi_param HTTPS on;
include fastcgi_params;
}
}
#配置lb01负载均衡
[root@lb01 conf.d]# vim lb.conf
upstream web {
server 172.16.1.7:80;
server 172.16.1.8:80;
server 172.16.1.9:80;
}
server {
listen 80;
server_name linux.discuz.com;
rewrite (.*) https://$server_name$1;
}
server {
listen 443 ssl;
server_name linux.discuz.com;
ssl_certificate /etc/nginx/ssl_key/server.crt;
ssl_certificate_key /etc/nginx/ssl_key/server.key;
location / {
proxy_pass http://web;
include proxy_params;
}
}
#配置host并访问
192.168.15.5 linux.discuz.com
创建数据库
[root@db01 ~]# yum install -y mariadb-server
[root@db01 ~]# systemctl enable --now mariadb
[root@db01 ~]# useradd www
[root@db01 ~]# mkdir /databases
[root@db01 ~]# chown -R www.www /databases/
[root@db01 ~]# mount -t nfs 172.16.1.31:/nfs/database /databases/
[root@db01 ~]# vim mysql_dump.sh
#!/bin/bash
DATE=`date +%F`
BACKUP="/databases"
cd $BACKUP
mysqldump -uroot -p123 --all-databases --single-transaction > mysql-all-${DATE}.sql
tar -czf mysql-all-${DATE}.tar.gz mysql-all-${DATE}.sql
rm -rf mysql-all-${DATE}.sql
[root@db01 ~]# chmod +x mysql_dump.sh
[root@db01 ~]# ./mysql_dump.sh
[root@db01 ~]# crontab -e
01 00 * * * /databases/mysql_dump.sh
[root@db01 ~]# mv mysql_dump.sh /databases/
创建数据库
[root@db01 ~]# mysqladmin -u root password '123'
登录数据库
[root@db01 ~]# mysql -uroot -p123
MariaDB [(none)]> show databases;
MariaDB [(none)]> create database discuz;
MariaDB [(none)]> grant all privileges on wecenter.* to www@'%''123';
MariaDB [(none)]> use mysql
MariaDB [(none)]> select host,user from user;
MariaDB [(none)]> drop database discuz;
[root@db01 ~]# systemctl restart mariadb
2597
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
