Nginx 代理 varnish 的姿势
TerHandfiel
8年前
<p>以前做网站的时候遇到了网站的访问量很大，而导致后端处理程序响应超时而导致的一些问题。当时采用的架构是nginx+php-fastcgi，同事想到了用nginx-proxycache来做页面缓存，效果也还行。下面我想介绍一下varnish的使用技巧</p> <h2><strong>准备</strong></h2> <p>varnish严格来说是可以当作一个代理服务器的软件，直接将 HTTP 请求转发到 php-cgi ,然后交给php处理,varnish会获取经过php处理后的数据,最后返回给浏览器。如图</p> <p style="text-align:center"><img src="https://simg.open-open.com/show/444380fc92f288794f187323c69dd10d.jpg"></p> <p>但是，现在 php-fastcgi 已经被逐渐淘汰了，也就是说我们一般情况下不会使用 php-fastcgi ，那么我们不能直接将varnish与php组合，因为php-fpm的交互方式为socket，而不再是监听本机的9000端口</p> <p>所以我们必须找一个的媒介,连接 varnish 和 php-fpm , nginx 可以扮演这个媒介,如下图：</p> <p style="text-align:center"><img src="https://simg.open-open.com/show/60071c1c861c29e6c1274502f785cf71.jpg"></p> <p>那么问题来了，根据研究发现， varnish 处理 http 请求不如 nginx 那么高效。所以如果我们让 nginx 做前锋，这样就更完美了。那我们需要怎么才能达到这个目的呢，下面我们来整理一下流程</p> <p style="text-align:center"><img src="https://simg.open-open.com/show/8d13afae02a727554b28811fd0b720bf.jpg"></p> <p>下面就来实现一下图三的架构吧。</p> <p>事先需要准备 nginx , varnish , php-fpm , php 这些软件，OS是 ubuntu ，所有软件都可以用 apt-get install 来安装，不了解包名全称的话可以先 apt-get update ，更新一下源，然后再用 apt-cache search xxx 来查找软件包名</p> <p>安装完 varnish 后，可以使用 service varnish 回车，查看可操作选项 * Usage: /etc/init.d/varnish {start|stop|restart|reload|force-reload|configtest} ,一般安装完毕后，系统会自动启动 varnish 的, nginx 也是一样,便不赘述了</p> <h2><strong>配置</strong></h2> <p>安装完所需的软件后，下面需要配置这些软件，来实现这个架构</p> <h3><strong>nginx部分</strong></h3> <p>vi /etc/nginx/nginx.conf</p> <pre> <code class="language-groovy">http { ## proxy global setting proxy_connect_timeout 5; proxy_read_timeout 60; proxy_send_timeout 5; proxy_buffer_size 16k; proxy_buffers 4 64k; proxy_busy_buffers_size 128k; ##END ## cache proxy pass upstream cache { server 127.0.0.1:6081; } ##END ## php proxy pass upstream php { server 127.0.0.1:8080; } ##END # Basic Settings sendfile on; tcp_nopush on; tcp_nodelay on; keepalive_timeout 65; types_hash_max_size 2048; server_tokens off; #depend on nginx-extras 需要安装nginx-extras才能定义Server more_set_headers 'Server: Bird-shark'; # server_names_hash_bucket_size 64; # server_name_in_redirect off; include /etc/nginx/mime.types; default_type application/octet-stream; ## # SSL Settings ## ssl_protocols TLSv1 TLSv1.1 TLSv1.2; # Dropping SSLv3, ref: POODLE ssl_prefer_server_ciphers on; ## # Logging Settings ## access_log /var/log/nginx/access.log; error_log /var/log/nginx/error.log; ## # Gzip Settings ## gzip on; gzip_disable "msie6"; gzip_vary on; gzip_proxied any; gzip_comp_level 6; gzip_buffers 16 8k; gzip_http_version 1.1; gzip_types text/plain text/css application/json application/javascript text/xml application/xml application/xml+rss text/javascript; ## # Virtual Host Configs ## include /etc/nginx/conf.d/*.conf; include /etc/nginx/sites-enabled/*; }</code></pre> <p>vi /etc/nginx/sites-available/default</p> <pre> <code class="language-groovy">server { listen 80 default_server; listen [::]:80 default_server; index index.html index.htm index.php; server_name localhost; location ~ .*\.(gif|jpg|png|css|js|flv|ico|swf|html)$ { proxy_set_header Host $host; proxy_set_header X-Real-IP $remote_addr; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; proxy_pass http://cache; } # pass the PHP scripts to FastCGI server listening on 127.0.0.1:9000 # location / { proxy_pass http://php; proxy_set_header Host $host; proxy_set_header X-Real-IP $remote_addr; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; proxy_pass_header Server; } } server { listen 8080; root /var/www/html; index index.html index.htm index.php; location / { if (!-e $request_filename){ rewrite ^(.*)$ /index.php?s=$1 last; break; } try_files $uri $uri/ =404; } location ~ ^(.+\.php)(.*)$ { fastcgi_pass unix:/var/run/php5-fpm.sock; fastcgi_intercept_errors on; fastcgi_buffers 8 128k; fastcgi_index index.php; fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name; include fastcgi_params; } }</code></pre> <h3><strong>varnish部分</strong></h3> <p>vi /etc/varnish/default.vcl</p> <pre> <code class="language-groovy">vcl 4.0; # Default backend definition. Set this to point to your content server. backend default { .host = "127.0.0.1"; .port = "8080"; } acl purgers { "localhost"; #"103.22.188.169"; } sub vcl_recv { # Happens before we check if we have this in cache already. # # Typically you clean up the request here, removing cookies you don't need, # rewriting the request, etc. if (req.restarts == 0) { unset req.http.X-Purger; } if(req.method == "PURGE"){ if(!client.ip ~ purgers){ return(synth(405,"Not Allowed.")); } return (purge); #ban("obj.http.x-url ~ " + req.url); } if(req.method == "GET" && req.url ~ "\.(js|css|jpg|png|gif|swf|jpeg|ico)$"){ unset req.http.cookie; } } sub vcl_backend_response { #set beresp.http.x-url = bereq.url; if (bereq.url ~ "\.(js|css|jpg|png|gif|swf|jpeg|ico)$") { unset beresp.http.Cache-Control; unset beresp.http.set-cookie; set beresp.ttl = 10h; set beresp.http.Cache-Control = "max-age=36000"; set beresp.do_gzip = true; } if(bereq.url ~ "\.html$"){ set beresp.ttl = 10m; set beresp.do_gzip = true; unset beresp.http.Cache-Control; unset beresp.http.Pragma; set beresp.http.Cache-Control = "max-age=600"; unset beresp.http.Expires; } } sub vcl_deliver { if (obj.hits > 0) { set resp.http.X-Cache = "HIT from " + req.http.host; #set resp.http.X-Cache-Hits = obj.hits; } else { set resp.http.X-Cache = "MISS from " + req.http.host; } if (req.http.X-Purger) { set resp.http.X-Purger = req.http.X-Purger; } unset resp.http.X-Powered-By; unset resp.http.Server; unset resp.http.Via; unset resp.http.X-Varnish; unset resp.http.Age; #unset resp.http.x-url; # Optional } sub vcl_hit { if (req.method == "PURGE") { return (synth(200,"Purged.")); } } sub vcl_miss { if (req.method == "PURGE") { return (synth(404,"Purged.")); } } sub vcl_purge { if (req.method == "PURGE") { #set req.http.X-Purge = "Purged"; ban("req.url ~ "+req.url); #return (restart); set req.method = "GET"; set req.http.X-Purger = "Purged"; return (restart); } } sub vcl_hash { hash_data(req.url); if (req.http.host) { hash_data(req.http.host); } else { hash_data(server.ip); } if (req.http.cookie) { hash_data(req.http.cookie); } if (req.http.Accept-Encoding ~ "gzip") { hash_data("gzip"); } elseif (req.http.Accept-Encoding ~ "deflate") { hash_data("deflate"); } }</code></pre> <h2><strong>测试&分析</strong></h2> <h3><strong>1. 在不使用缓存模块的情况下</strong></h3> <p>vi /etc/nginx/sites-available/default</p> <pre> <code class="language-groovy">#location ~ .*\.(gif|jpg|png|css|js|flv|ico|swf|html)$ { # proxy_set_header Host $host; # proxy_set_header X-Real-IP $remote_addr; # proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; # proxy_pass http://cache; #}</code></pre> <p>先用chrome浏览器访问查看请求头</p> <p><img src="https://simg.open-open.com/show/4d8782a013b6b091325970266f828cb4.png"></p> <p>我们再使用curl,在服务器上执行以下命令</p> <pre> <code class="language-groovy">curl -k -v 'http://192.168.99.1/Public/Home/images/t_navigation_logo.png' -H 'Pragma: no-cache' -H 'Accept-Encoding: gzip, deflate, sdch' -H 'Accept-Language: zh,en;q=0.8,zh-CN;q=0.6' -H 'Upgrade-Insecure-Requests: 1' -H 'User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/52.0.2743.116 Safari/537.36' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8' -H 'Connection: keep-alive' --compressed</code></pre> <p>发现有输出内容。</p> <p>然后，反选disable cache</p> <p><img src="https://simg.open-open.com/show/b72c9b38f0b016c9dd1d14faf1f73745.png"></p> <p>然后在服务器上执行以下命令</p> <pre> <code class="language-groovy">curl -k -v 'http://192.168.99.1/Public/Home/images/t_navigation_logo.png' -H 'If-None-Match: "57c6b733-1962"' -H 'Accept-Encoding: gzip, deflate, sdch' -H 'Accept-Language: zh,en;q=0.8,zh-CN;q=0.6' -H 'Upgrade-Insecure-Requests: 1' -H 'User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/52.0.2743.116 Safari/537.36' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8' -H 'Connection: keep-alive' -H 'If-Modified-Since: Wed, 31 Aug 2016 10:53:39 GMT' --compressed</code></pre> <p>发现只返回了头部信息，然而没有内容返回</p> <p>然后我们比较两个命令 发现区别就在 -H 'Pragma: no-cache' 和 -H 'If-Modified-Since: Wed, 31 Aug 2016 10:53:39 GMT' -H 'If-None-Match: "57c6b733-1962"'</p> <p>57c6b733-1962 这串字符对应的是服务器响应给浏览器的 ETag 部分的内容，然后我们修改一下部分的内容</p> <pre> <code class="language-groovy">curl -k -v 'http://192.168.99.1/Public/Home/images/t_navigation_logo.png' -H 'If-None-Match: "57c6b733-1234"' -H 'Accept-Encoding: gzip, deflate, sdch' -H 'Accept-Language: zh,en;q=0.8,zh-CN;q=0.6' -H 'Upgrade-Insecure-Requests: 1' -H 'User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/52.0.2743.116 Safari/537.36' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8' -H 'Connection: keep-alive' -H 'If-Modified-Since: Wed, 31 Aug 2016 10:53:39 GMT' --compressed</code></pre> <p>在服务器端执行一下。发现有内容返回，所以这个 ETag 相当于 token ，它不是由 nginx 随便生成的，且跟请求链接应是一一对应的，用来标识缓存的,当服务器返回的状态为 304 的时候,这时候我们浏览器会直接找到本地的缓存数据</p> <h3><strong>2. 在使用缓存模块的情况下</strong></h3> <p>vi /etc/nginx/sites-available/default</p> <pre> <code class="language-groovy">location ~ .*\.(gif|jpg|png|css|js|flv|ico|swf|html)$ { proxy_set_header Host $host; proxy_set_header X-Real-IP $remote_addr; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; proxy_pass http://cache; }</code></pre> <p>用浏览器查看响应头</p> <p>发现 X-Cache:MISS from 192.168.99.1 .这表示缓存未命中，然后我们刷新 X-Cache:HIT from 192.168.99.1 ，这时候发现已经命中了。</p> <p>对于已经命中的资源文件，我们如果将其删除会出现什么效果呢，答案是，其依然可以访问，除非重启或者将缓存清除</p> <p>但是对PURGE显然是不对外公开的,以下是服务器端用curl清除varnish缓存的命令</p> <pre> <code class="language-groovy">curl -v -k -X PURGE http://localhost/Public/Home/css/t_navigation.css</code></pre> <h2><strong>结语</strong></h2> <p>varnish 是一款内存类型的缓存软件，而非nginx扩展proxy_cache那种物理缓存类型的软件，存取速度比较快,但是也有弊端，重启后所有缓存得重写。不管怎么说，什么架子都适用的场景，要想满足业务需求还是得捣鼓透彻，而我也只是将我想到的给实现出来，毕竟资源和精力都是有限的，也就随便玩玩，诸位看客看看就好，别太认真，知道怎么回事儿就行。</p> <p> </p> <p>来自：https://segmentfault.com/a/1190000006853425</p> <p> </p>