高可用集群 HA high availability
避免单节点故障
软件:keepalived
负载均衡集群 LB load balance
提高负载,提高并发量
软件:nginx反向代理 lvs
硬件负载均衡器 F5(BigIP)和redware
HPC高性能运算集群
分布式存储集群
极大的提升存储容量,提供数据高可用,保证数据安全
软件:ceph
Nginx proxy 是 Nginx 的王牌功能,利用 proxy 基本可以实现一个完整的 7 层负载均。
1. 功能强大,性能卓越,运行稳定。
2. 配置简单灵活。
3. 能够自动剔除工作不正常的后端服务器。
4. 上传文件使用异步模式。
5. 支持多种分配策略,可以分配权重,分配方式灵活。
如果有tomcat,apache,squid 设置为如下:
#vim /usr/local/nginx/conf/nginx.conf # 在最后添加以下内容,定义服务器组
upstream tomcat_servers{拓扑
++++++++++++
+ Client + 192.168.122.1/24 (真实机做客户端)
++++++++++++
|
|
++++++++++++ 192.168.122.254/24
+ Nginx +
++++++++++++
|
__________________________|____________________________
________________|________________ ___________|__________________
| | | |
++++++++++++ ++++++++++++ ++++++++++++ ++++++++++++
+ HTML A + + HTML B + + PHP A + + PHP B +
++++++++++++ ++++++++++++ ++++++++++++ ++++++++++++
eth0 192.168.122.10/24 eth0 192.168.122.20/24 eth0 192.168.122.30/24 eth0 192.168.122.40/24
nginx dr director
|
--------------------
| |
nginx nginx rs realserver
给realserver安装nginx,作页面,启动服务,保证能正常访问
HTML A & HTML B
[root@localhost ~]# yum install httpd
分别创建测试页面 index.html ,开启服务
PHP A & php B
[root@localhost ~]# yum install httpd
分别创建测试页面 index.php ,开启服务
安装配置Nginx
[root@localhost ~]# rpm -ivh nginx-0.6.36-1.el5.i386.rpm
[root@localhost ~]# vim /etc/nginx/nginx.conf
location / {
root /usr/share/nginx/html;
index index.html index.htm;
if ($request_uri ~* \.html$) {
proxy_pass http://htmlserver;
}
if ($request_uri ~* \.php$) {
proxy_pass http://phpserver;
}
}
[root@localhost ~]# vim /etc/nginx/conf.d/test.conf
upstream htmlserver {
server 192.168.122.10;
server 192.168.122.20;
}
upstream phpserver {
server 192.168.122.30;
server 192.168.122.40;
}
[root@localhost ~]# service nginx start
在客户端访问 Nginx 测试
[root@localhost ~]# elinks –dump http:// 192.168.122.254
[root@localhost ~]# elinks –dump http:// 192.168.122.254/index.html
[root@localhost ~]# elinks –dump http:// 192.168.122.254/index.php
===============================
upstream支持的负载均衡算法(面试题)
轮询(默认): 可以通过weight指定轮询的权重,权重越大,被调度的次数越多 rr round robin
权重:用数字 谁数字大谁权重就高 按比例 1 3
rr
wrr
ip_hash: 根据每个请求IP进行调度,可以解决session的问题,不能使用weight
client_ip 192.168.1.8 nginx反向 webserver1
fair: 可以根据请求页面的大小和加载时间长短进行调度,使用第三方的upstream_fair模块
url_hash: 按请求的url的hash进行调度,从而使每个url定向到同一服务器,使用第三方的hash模块
upstream支持的状态参数
down: 暂停对该服务器的调度
backup: 类似于LVS Sorry Server,当所有的非backup的服务器故障
max_fails: 请求失败的次数,默认为1
fail_timeout: 在经历max_fails次失败后,暂停服务的时间
upstream tianyun.com {
# ip_hash;
server 192.168.10.137 weight=1 max_fails=2 fail_timeout=2;
server 192.168.10.20 weight=2 max_fails=2 fail_timeout=2;
server 192.168.10.251 max_fails=2 fail_timeout=5 down;
server 192.168.10.253 backup;
}
注:当使用ip_hash时,服务器状态不可使用weight和backup
=================================================================================
Nginx实现七层的负载均衡
调度到同一组上游服务器
=================================================================================
拓扑结构
[LB Nginx]
192.168.1.2
[httpd] [httpd] [httpd]
192.168.1.3 192.168.1.4 192.168.1.5
实施过程
1. nginx
http {
upstream httpservers {
server 192.168.1.3:80 weight=1 max_fails=2 fail_timeout=2;
server 192.168.1.4:80 weight=2 max_fails=2 fail_timeout=2;
server 192.168.1.5:80 weight=2 max_fails=2 fail_timeout=2;
server 192.168.1.100:80 backup;
}
location / {
proxy_pass http://httpservers;
proxy_next_upstream error timeout invalid_header http_500 http_502 http_503 http_504;
}
}
proxy_next_upstream:这个指令属于 http_proxy 模块的,指定后端返回什么样的异常响应时,使用另一个realserver
2. Apache LogFormat 可选
LogFormat "%{X-Real-IP}i %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"" combined
LogFormat "%h %l %u %t \"%r\" %>s %b" common
LogFormat "%{Referer}i -> %U" referer
LogFormat "%{User-agent}i" agent
=================================================================================
Nginx实现七层的负载均衡
调度到不同组上游服务器
1. 动静分离
2. 网站进行分区
=================================================================================
拓扑结构
[vip: 192.168.1.80]
[LB1 Nginx] [LB2 Nginx]
192.168.1.2 192.168.1.3
[news] [milis] [videos] [images] [others]
1.11 1.21 1.31 1.41 1.51
1.12 1.22 1.32 1.42 1.52
1.13 1.23 1.33 1.43 1.53
... ... ... ... ...
一、实施过程
1. 根据站点分区进行调度
http {
upstream news {
server 192.168.1.11:80 weight=1 max_fails=2 fail_timeout=2;
server 192.168.1.12:80 weight=2 max_fails=2 fail_timeout=2;
server 192.168.1.13:80 weight=2 max_fails=2 fail_timeout=2;
}
upstream milis {
server 192.168.1.21:80 weight=1 max_fails=2 fail_timeout=2;
server 192.168.1.22:80 weight=2 max_fails=2 fail_timeout=2;
server 192.168.1.23:80 weight=2 max_fails=2 fail_timeout=2;
}
upstream videos {
server 192.168.1.31:80 weight=1 max_fails=2 fail_timeout=2;
server 192.168.1.32:80 weight=2 max_fails=2 fail_timeout=2;
server 192.168.1.33:80 weight=2 max_fails=2 fail_timeout=2;
}
upstream images {
server 192.168.1.41:80 weight=1 max_fails=2 fail_timeout=2;
server 192.168.1.42:80 weight=2 max_fails=2 fail_timeout=2;
server 192.168.1.43:80 weight=2 max_fails=2 fail_timeout=2;
}
upstream others {
server 192.168.1.51:80 weight=1 max_fails=2 fail_timeout=2;
server 192.168.1.52:80 weight=2 max_fails=2 fail_timeout=2;
server 192.168.1.53:80 weight=2 max_fails=2 fail_timeout=2;
}
server {
location / {
proxy_pass http://others;
}
location /news {
proxy_pass http://news;
}
location /mili {
proxy_pass http://milis;
}
location ~* \.(wmv|mp4|rmvb)$ {
proxy_pass http://videos;
}
location ~* \.(png|gif|jpg)$ {
proxy_pass http://images;
}
}
2. 根据动静分离进行调度
http {
upstream htmlservers {
server 192.168.1.3:80 weight=1 max_fails=2 fail_timeout=2;
server 192.168.1.4:80 weight=2 max_fails=2 fail_timeout=2;
}
upstream phpservers {
server 192.168.1.3:80 weight=1 max_fails=2 fail_timeout=2;
server 192.168.1.4:80 weight=2 max_fails=2 fail_timeout=2;
}
server {
location ~* \.html$ {
proxy_pass http://htmlservers;
}
location ~* \.php$ {
proxy_pass http://phpservers;
}
}
}
二、Keepalived实现调度器HA
注:主/备调度器均能够实现正常调度
1. 主/备调度器安装软件
[root@master ~]# yum -y install ipvsadm keepalived
[root@backup ~]# yum -y install ipvsadm keepalived
2. Keepalived
Master
# vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
router_id director1 //辅助改为director2
}
vrrp_instance VI_1 {
state BACKUP
nopreempt
interface eth0 //心跳接口,尽量单独连接心跳
virtual_router_id 80 //MASTER,BACKUP一致
priority 100 //辅助改为50
advert_int 1
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.80
}
}
BACKUP
3. 启动KeepAlived(主备均启动)
# chkconfig keepalived on
# service keepalived start
# ip addr
4. 扩展对调度器Nginx健康检查(可选)
思路:
让Keepalived以一定时间间隔执行一个外部脚本,脚本的功能是当Nginx失败,则关闭本机的Keepalived
a. script
[root@master ~]# cat /etc/keepalived/check_nginx_status.sh
#!/bin/bash
/usr/bin/curl -I http://localhost &>/dev/null
if [ $? -ne 0 ];then
/etc/init.d/keepalived stop
fi
[root@master ~]# chmod a+x /etc/keepalived/check_nginx_status.sh
b. keepalived使用script
! Configuration File for keepalived
global_defs {
router_id director1
}
vrrp_script check_nginx {
script "/etc/keepalived/check_nginx_status.sh"
interval 5
}
vrrp_instance VI_1 {
state BACKUP
interface eth0
nopreempt
virtual_router_id 90
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass tianyun
}
virtual_ipaddress {
192.168.1.80
}
track_script {
check_nginx
}
}
--------------------
proxy_pass
后端服务器用的非php独立进程
apache+php模块
fastcgi_pass
后端服务器用的是php-fpm
php-fpm(fastcgi形式的php)
后端服务器部署详细过程:
安装软件:
# yum install nginx php php-fpm -y
# vim /etc/nginx/nginx.conf //添加php配置
在server里面添加如下配置:
location ~ \.php$ {
root html;
fastcgi_pass 127.0.0.1:9000;
fastcgi_index index.php;
fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
include fastcgi_params;
}
修改php-fpm进程账户并开启php-fpm的进程: 端口是9000
#vim /etc/php-fpm.d/www.conf //修改如下参数,默认值是apache
user = nginx
group = nginx
为什么设置成nginx:
因为nginx.conf配置的账户为nginx
# systemctl start php-fpm
前端nginx反向代理服务器:
upstream web {
server 10.0.0.21;
server 10.0.0.22;
}
upstream phpserver {
server 10.0.0.23;
server 10.0.0.24;
} #上面的配置写到http里面server外面
server {
listen 80;
server_name www.baidu.com;
location / { #html的配置
proxy_pass http://web;
}
location ~* \.php$ { #php的配置
proxy_pass http://phpserver;
}
#yum install epel-release -y
#yum nginx php php-fpm -y
php-fpm是独立的php进程 端口:9000