●VCL 中文man page(3)

发布时间:2019-07-15 10:49:39编辑:auto阅读(1281)

             EXAMPLES(例子)

                      下面这段代码和默认的配置相同,后端服务器主机名设置为“backend.exampl.com
                       backend default {

                        .host = "backend.example.com";

                        .port = "http";

                       }

     
                       sub vcl_recv {

                         if (req.http.x-forwarded-for) {

                         set req.http.X-Forwarded-For = req.http.X-Forwarded-For ", " client.ip;

                       } else {

                         set req.http.X-Forwarded-For = client.ip;

                       }

     
                       if (req.request != "GET" &&

                       req.request != "HEAD" &&

                       req.request != "PUT" &&

                       req.request != "POST" &&

                       req.request != "TRACE" &&

                       req.request != "OPTIONS" &&

                       req.request != "DELETE") {

                         // Non-RFC2616 or CONNECT which is weird.

                         return (pipe);

                         }

                       if (req.request != "GET" && req.request != "HEAD") {

                       // We only deal with GET and HEAD by default

                         return (pass);

                         }

                       if (req.http.Authorization || req.http.Cookie) {

                         // Not cacheable by default

                         return (pass);

                         }

                       return (lookup);

                       }

     
                       sub vcl_pipe {

                         # Note that only the first request to the backend will have

                         # X-Forwarded-For set.  If you use X-Forwarded-For and want to

                         # have it set for all requests, make sure to have:

                         # set req.http.connection = "close";

                         # here.  It is not set by default as it might break some broken web

                         # applications, like IIS with NTLM authentication.

                         return (pipe);

                       }

     
                       sub vcl_pass {

                         return (pass);

                       }

     
                       sub vcl_hash {

                         set req.hash += req.url;

                         if (req.http.host) {

                           set req.hash += req.http.host;

                         } else {

                           set req.hash += server.ip;

                         }

                         return (hash);

                       }

     
                       sub vcl_hit {

                         if (!obj.cacheable) {

                         return (pass);

                         }

                       return (deliver);

                       }

     
                       sub vcl_miss {

                         return (fetch);

                       }

     
                       sub vcl_fetch {

                       if (!beresp.cacheable) {

                         return (pass);

                       }

                       if (beresp.http.Set-Cookie) {

                         return (pass);

                         }

                       return (deliver);

                       }

     
                       sub vcl_deliver {

                         return (deliver);

                       }

     
                       sub vcl_error {

                       set obj.http.Content-Type = "text/html; charset=utf-8";

                       synthetic {"

                       <?xml version="1.0" encoding="utf-8"?>

                       <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"

                       "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">

                       <html>

                       <head>

                       <title>"} obj.status " " obj.response {"</title>

                       </head>

                       <body>

                       <h1>Error "} obj.status " " obj.response {"</h1>

                       <p>"} obj.response {"</p>

                       <h3>Guru Meditation:</h3>

                       <p>XID: "} req.xid {"</p>

                       <hr>

                       Varnish cache server

                       </body>

                       </html>

                       "};

                       return (deliver);

                       }

             下面的例子显示一个varnishd实例支持多个独立的站点,基于请求的URL选择使用的后端服务器:
                       backend www {

                         .host = "www.example.com";

                         .port = "80";

                       }

     
                       backend p_w_picpaths {

                         .host = "p_w_picpaths.example.com";

                         .port = "80";

                       }

     
                       sub vcl_recv {

                         if (req.http.host ~ "^(www.)?example.com$") {

                           set req.http.host = "www.example.com";

                           set req.backend = www;

                         } elsif (req.http.host ~ "^p_w_picpaths.example.com$") {

                           set req.backend = p_w_picpaths;

                         } else {

                           error 404 "Unknown virtual host";

                         }

                       }

     
                       The following snippet demonstrates how to force a minimum TTL for

                       all documents.  Note that this is not the same as setting the

                       default_ttl run-time parameter, as that only affects document for

                       which the backend did not specify a TTL:::

     
                       sub vcl_fetch {

                         if (obj.ttl < 120s) {

                         set obj.ttl = 120s;

                         }

                       }

             下面这段代码用来强制缓存带cookies的内容:
                       sub vcl_recv {

                         if (req.request == "GET" && req.http.cookie) {

                         call(lookup);

                         }

                       }

     
                       sub vcl_fetch {

                         if (beresp.http.Set-Cookie) {

                         deliver;

                        }

                       }

             下面代码的作用是利用squidHTTP PURGE模式清理无法使用的目标。
                       acl purge {

                         "localhost";

                         "192.0.2.1"/24;

                       }

     
                       sub vcl_recv {

                         if (req.request == "PURGE") {

                         if (!client.ip ~ purge) {

                           error 405 "Not allowed.";

                         }

                         lookup;

                         }

                       }

     
                       sub vcl_hit {

                         if (req.request == "PURGE") {

                           set obj.ttl = 0s;

                           error 200 "Purged.";

                         }

                       }

     
                       sub vcl_miss {

                         if (req.request == "PURGE") {

                         error 404 "Not in cache.";

                         }

                       }

             SEE  ALSO

                       Varnishd1
            
             HISTORY

                       The VCL language was developed by Poul-Henning Kamp in cooperation with Verdens       Gang AS, Linpro AS and Varnish Software. This manual page was written by Dag-Erling         Smørgrav and later edited by Poul-Henning Kamp and Per Buer.
            
             COPYRIGHT

             这个文档的版权和varnish自身的版本一样,请看LICENCE
                           * Copyright (c) 2006 Verdens Gang AS
                             * Copyright (c) 2006-2008 Linpro AS
                           * Copyright (c) 2008-2010 Redpill Linpro AS
                           * Copyright (c) 2010 Varnish Software AS

关键字