大型網站架構之百萬PV

• 簡介

PV即點擊量，通常是衡量一個網絡新聞頻道或網站甚至一條網絡新聞的重要指標。PV從某種程度上已經成爲投資者衡量商業網站表現的最重要的尺度。

PV是一個訪問者在24小時到底看了網站的幾個頁面。

• 案例描述

本案例設計採用四層模式實現，主要分爲前端反向代理，web層，數據庫緩存層和數據庫層。前端反向代理層採用主備模式，web層採用集羣模式，數據庫緩存採用主備模式，數據庫層採用主從模式。

• 案例環境

主：192.168.177.145 centos7-1

從：192.168.177.135 centos7-2

節點1：192.168.177.132 centos7-3

節點2：192.168.177.133 centos7-4

主服務器 從服務器

• 安裝帶有nginx rpm軟件包的源

  # rpm -ivh http://nginx.org/packages/centos/7/noarch/RPMS/\
  nginx-release-centos-7-0.el7.ngx.noarch.rpm

• 使用centos 默認倉庫完成下面的安裝 主

  #  yum install -y keepalived nginx
  #  vim /etc/keepalived/keepalived.conf  //從上修改三個參數

  ! Configuration File for keepalived

vrrp_script nginx {
script "/opt/shell/nginx.sh"
interval 2
} //添加
global_defs {
route_id NGINX_HA //修改
}
vrrp_instance VI_1 {
state MASTER
interface ens33
virtual_router_id 51
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
track_script {
nginx
}
virtual_ipaddress {
192.168.177.188 //虛擬IP
192.168.200.188
}
}

#  mkdir /opt/shell  
#  cd /opt/shell
#  vim /opt/shell/nginx.sh

#!/bin/bash
k=`ps -ef | grep keepalived | grep -v grep | wc -l`
if [ $k -gt 0 ];then
    /bin/systemctl start nginx.service
else
/bin/systemctl stop nginx.service
fi

# chmod +x /opt/shell/nginx.sh   //變成可執行的腳本

從服務器與主服務器配置一樣（內容要修改）

# yum install -y keepalived nginx
# vim /etc/keepalived/keepalived.conf  //從上修改三個參數

! Configuration File for keepalived

vrrp_script nginx {
    script "/opt/shell/nginx.sh"
    interval 2
}    //添加
global_defs {
    route_id NGINX_HB   //修改
}
vrrp_instance VI_1 {
    state MASTER
    interface ens33
    virtual_router_id 52
    priority 90
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
}
track_script {
    nginx
}
virtual_ipaddress {
    192.168.177.188      //虛擬IP
    192.168.200.188
    }
}
# mkdir /opt/shell
# cd /opt/shell
# vim /opt/shell/nginx.sh

#!/bin/bash
k=`ps -ef | grep keepalived | grep -v grep | wc -l`
if [ $k -gt 0 ];then
    /bin/systemctl start nginx.service
else
/bin/systemctl stop nginx.service
fi

# chmod +x /opt/shell/nginx.sh   //變成可執行的腳本

配置nginx前端調度功能(主，從)

# vim /etc/nginx/nginx.conf //在#gzip on;下面添加：

 upstream tomcat_pool {
                server 192.168.177.132:8080;
                server 192.168.177.133:8080;
                ip_hash;           #會話穩固功能，否則無法通過vip地址登陸
        }
        server {
                listen 80;
                server_name 192.168.177.188; #虛擬出的IP  
                location / {
                        proxy_pass http://tomcat_pool;
                        proxy_set_header X-Real-IP $remote_addr;
                }
    }

# nginx -t -c /etc/nginx/nginx.conf  //測試配置文件語法
# systemctl start keepalived.service    //nginx啓動會等待一會

兩臺節點都要做tomcat

# tar xf apache-tomcat-8.5.23.tar.gz
# tar xf jdk-8u144-linux-x64.tar.gz

# mv apache-tomcat-8.5.23/ /usr/local/tomcat8
# mv jdk1.8.0_144/ /usr/local/java

# vim /etc/profile

export JAVA_HOME=/usr/local/java
export JRE_HOME=/usr/local/java/jre
export PATH=$PATH:/usr/local/java/bin
export CLASSPATH=./:/usr/local/java/lib:/usr/local/java/jre/lib

# source /etc/profile

# java -version

# ln -s /usr/local/tomcat8/bin/startup.sh /usr/bin/tomcatup
# ln -s /usr/local/tomcat8/bin/shutdown.sh /usr/bin/tomcatdown

# tomcatup     //開啓服務
# netstat -anpt | grep 8080

# vim /usr/local/tomcat8/webapps/ROOT/index.jsp //修改默認網頁內容

<h1>Server 132!!!</h1>
http://192.168.177.132:8080/         //測試默認測試頁是否正常顯示（節點）
http://192.168.177.133:8080/
http://192.168.177.188/  輸入調度器地址，也就是虛擬地址，測試兩臺節點的調度情況。

將192.168.177.132宕機（tomcatdown)，看看http://192.168.177.188/會不會顯示<h1>Server 133!!!</h1>

# vim server.xml //跳到行尾，在Host name下新增 148行

<Context path="" docBase="SLSaleSystem" reloadable="true" debug="0"></Context>
日誌調試信息debug爲0表示信息越少，docBase指定訪問目錄

Mysql安裝（主，從）

# yum install -y mariadb-server mariadb
# systemctl start mariadb.service
# systemctl enable mariadb.service    //開機自啓動
# netstat -anpt | grep 3306
# mysql_secure_installation    //常規安全設置
# mysql -u root -p

導入數據庫

先掛載
# mysql -u root -p  < slsaledb-2014-4-10.sql
# mysql -u root -p
  show databases;
  GRANT all ON slsaledb.* TO 'root'@'%' IDENTIFIED BY 'abc123';
  flush privileges;

以下在兩臺tomcat節點做

# tar xf SLSaleSystem.tar.gz -C /usr/local/tomcat8/webapps/
# cd /usr/local/tomcat8/webapps/SLSaleSystem/WEB-INF/classes
# vim jdbc.properties //修改數據庫IP地址是VRRP的虛擬IP，以及授權的用戶名root和密碼abc123。

網站測試

http://192.168.177.132:8080/   //默認的用戶名admin 密碼：123456
http://192.168.177.133:8080/

http://192.168.177.188  //輸入虛擬地址測試登錄，並且關閉主再測試登錄

redis集羣（主，從）

# yum install -y epel-release
# yum install redis -y
# vim /etc/redis.conf
  bind 0.0.0.0
# systemctl start redis.service
# netstat -anpt | grep 6379
# redis-cli -h 192.168.177.145 -p 6379 //測試連接
  192.168.177.145:6379> set name test  //設置name 值是test
  192.168.177.145:6379> get name //獲取name值

從服務器上多如下一行配置

 266行 slaveof 192.168.177.145  6379 //主服務器的IP不是虛擬IP

# redis-cli -h 192.168.177.135 -p 6379 //登錄從，獲取值，成功說明主從同步成功
  192.168.177.135:6379> get name
"test"

配置商城項目中連接redis的參數

38         <!--redis 配置 開始-->

  47                 <constructor-arg value="192.168.177.188"/>
  48                 <constructor-arg value="6379"/>

測試緩存效果

# redis-cli -h 192.168.177.188 -p 6379
  192.168.177.188:6379> info
 keyspace_hits:1  或者 keyspace_misses:2//關注這個值，命中數和未命中數

登錄商城，然後反覆點擊需要數據庫參與的操作頁面，再回來檢查keyspace_hits或者keyspace_misses: 值變化。

配置redis集羣主從切換，只在主服務器上操作

# redis-cli -h  192.168.177.145 info Replication       //獲取當前服務器的角色
# vim /etc/redis-sentinel.conf
17行 protected-mode no17G
68行 sentinel monitor mymaster 192.168.177.145 6379 1 //1表示1臺從 注意：修改
98行 sentinel down-after-milliseconds mymaster 3000 //故障切換時間單位是毫秒
# service redis-sentinel start //啓動集羣
# netstat -anpt | grep 26379
# redis-cli -h 192.168.177.145 -p 26379 info Sentinel //查看集羣信息

驗證主從切換

在主上
# systemctl stop redis.service
# redis-cli -h 192.168.177.145 -p 26379 info Sentinel //發現主變成了135

驗證數據同步情況

# redis-cli -h 192.168.177.135 -p 6379 
  192.168.177.135:6379> set name2 test2
OK
  192.168.177.135:6379> get name2
"test2"
# service redis start //把主啓動
# redis-cli -h 192.168.177.145 -p 6379
  192.168.175.128:6379> get name2
  "test2"

Mysql主從

mysql主服務器配置

# vim /etc/my.cnf //[mysqld]下

binlog-ignore-db=mysql,information_schema
character_set_server=utf8
log_bin=mysql_bin
server_id=1
log_slave_updates=true
sync_binlog=1

# systemctl restart mariadb
# netstat -anpt | grep 3306
# mysql -u root -p
  show master status; //記錄日誌文件名稱和 位置值
  grant replication slave on *.* to 'rep'@'192.168.177.%' identified by '123456';
  flush privileges;

mysql從服務器配置

# vim /etc/my.cnf //[mysqld]下

binlog-ignore-db=mysql,information_schema
character_set_server=utf8
log_bin=mysql_bin
server_id=2
log_slave_updates=true
sync_binlog=1

# systemctl restart mariadb
# netstat -anpt | grep 3306
# mysql -u root -p

  change master to master_host='192.168.177.145',master_user='rep',master_password='123456',master_log_file='mysql_bin.000001',master_log_pos=245;

  start slave;

  show slave status\G;

    Slave_IO_Running: Yes
    Slave_SQL_Running: Yes