CDN搭建

CDN

client->cdn(squid->HA+F5/LVS(4)->HA+nginx/haproxy(7)->web(apache/nginx)
->app(php/jsp)->DB(mysql&redis)

zabbix openstack kvm/docker/k8s|swarm   hadoop

nginx可以反向代理
HA高可用
nginx(异步非阻塞)
apache(同步非阻塞)
LVS(linux virtual server)                 

 

搭建过程
一 设置母盘
最小化安装rhel6.5
1关闭selinux
2搭建yum源
真机搭建http源
mkdir -p /var/www/html/rhel6.5
mount /root/rhel-server-6.5-x86_64-dvd.iso /var/www/html/rhel6.5/

yum install openssh lftp -y
3关闭ip6  iptables
/etc/init.d/iptables stop
/etc/init.d/ip6tables stop
chkconfig iptables off
chkconfig ip6tables off
4删掉绑定的网卡MAC地址
rm -fr /etc/udev/rules.d/70-persistent-net.rules

创建快照
qemu-img create -f qcow2 -b /var/lib/libvirt/images/rhel6.5母盘.qcow2   /var/lib/libvirt/images/server1.qcow2
server2
server3

修改主机名
vi /etc/sysconfig/network
server1
server2
server3

 

server1虚拟机作为使用varnish服务器
下载varnish包
mget varnish-*
库文件也要安装
在安装包目录下
yum install varnish-* -y
查看配置文件
rpm -qc varnish-.0

首先查看一下启动的要求，不需要修改这个文件
vim /etc/sysconfig/varnish
NFILES 最大打开文件个数
MEMLOCK=82000  默认是运行varnish需要82M内存
VARNISH_LISTEN_PORT=80  端口改为80端口

sysctl -a |grep file
查看自己系统最大打开文件数
free -m 查看内存大小

改系统参数
vi /etc/security/limits.conf
末尾增加
varnish      -    nofile   131072  
varnish      -    memlock  82000
varnish      -    nproc    unlimited

vi /etc/varnish/default.vcl
backend default {
  .host = "172.25.11.2";
  .port = "80";
}

然后启动 /etc/init.d/varnish start
效果是 curl访问server1时候自动加载server2

现在写入判断
backend default {
  .host = "172.25.11.2";
  .port = "80";
}
sub vcl_deliver{
if(obj.hits>0){
set resp.http.X-Cache = "HIT from HZP cache";
}
else{
set  resp.http.X-Cache="MISS from HZP cache";
}
return (deliver);
}
}

清全部缓存  varnishadm ban.url .*$   
清某个页面    varnishadm ban.url    /index.html

其中sub vcl_deliver函数的意思是

如果缓存中有，就返回HIT  不然就返回MISS 然后再去服务器中缓存

再次编辑

backend web1 {
  .host = "172.25.11.2";
  .port = "80";
}
backend web2 {
  .host = "172.25.11.3";
  .port = "80";
}
sub vcl_recv {
   if (req.http.host ~ "^(www.)?westos.org") {
        set req.http.host = "www.westos.org";
        set req.backend = web1;
} elsif (req.http.host ~ "^bbs.westos.org") {
        set req.backend = web2;
} else {
        error 404 "hzp cache else";
        }
}

sub vcl_deliver{
if(obj.hits>0){
set resp.http.X-Cache = "HIT from HZP cache";
}
else{
set  resp.http.X-Cache="MISS from HZP cache";
}
return (deliver);
}

###其中sub vcl_recv函数意思 ：

如果 输入www.westos.com 就返回web1的地址的http内容

bbs则返回web2的内容

/etc/init.d/varnish reload
curl 172.25.11.1 -I

 

负载均衡
backend web1 {
  .host = "172.25.11.2";
  .port = "80";
}
backend web2 {
  .host = "172.25.11.3";
  .port = "80";
}

director lb round-robin {
        {.backend=web1;}
        {.backend=web2;}
}

sub vcl_recv {
   if (req.http.host ~ "^(www.)?westos.org") {
        set req.http.host = "www.westos.org";
        set req.backend = lb;
                return (pass);   #为了测试加入的,实际上删掉
} elsif (req.http.host ~ "^bbs.westos.org") {
        set req.backend = web2;
}       else{
        error 404 "westos cache";}
}
sub vcl_deliver{
if(obj.hits>0){
set resp.http.X-Cache = "HIT from HZP cache";
}
else{
set  resp.http.X-Cache="MISS from HZP cache";
}
return (deliver);
}

server1中 varnishadm ban.url .*$

真机访问bbs.westos.com  几次就会均衡