关于nginx中的rewrite,之前的理解总感觉有些不精确。比如以下问题,经过rewrite之后:
什么情况会返回200?
什么情况会返回301/302?
什么情况浏览器里的url不变?
什么情况浏览器里的url会变?
什么情况匹配一次就不再匹配之后的规则或location?
什么情况匹配到一条规则后,会以rewrite之后的url再到server段走一遍
如果读者能理解清楚以上问题,则说明对nginx的rewrite整体上已经有了全面的认识。
接下来我们针对以上问题,来通过试验逐一解答
实验环境:nginx 1.8
网站根目录:nginx/html
注:在http段或者server段添加rewrite_log on;并且设置error_log为notice级别,可以在error og里捕捉到rewrite的过程
一、先实验一个最简单的:http://172.16.25.162/ljk.html。该文件不在在html下,而在在html/wordpress下
server { listen 80; server_name localhost; rewrite_log on; location / { root /usr/local/nginx/html; index index.shtml index.php index.html; if (!-e $request_filename) { rewrite ^/(.+\..+)$ /wordpress/$1; } } 观察nginx访问日志和错误日志 access_log: "GET /ljk.html HTTP/1.1" 200 error_log: *98 "^/(.+\..+)$" matches "/ljk.html", client: 172.16.25.3, server: localhost, request: "GET /ljk.html *98 rewritten data: "/wordpress/ljk.html", args: "", client:172.16.25.3, server: localhost, request: "GET /ljk.html 浏览器里的url没有改变,状态码返回200 由此可见,如此配置实现了一个最简单的内部跳转。 |
二、再进一步,看一个两级的rewrite。将html/wordpress/下的ljk.html移到html/ljk/下,并且在location /wordpress 下配置重写规则
location / { root /usr/local/nginx/html; index index.shtml index.php index.html; if (!-e $request_filename) { rewrite ^/(.+\..+)$ /wordpress/$1; } } location /wordpress { if (!-e $request_filename) { rewrite ^/wordpress/(.+\..+)$ /ljk/$1; } } access_log: "GET /ljk.html HTTP/1.1" 200 error_log:可以看到两次匹配与重写的过程 *99 "^/(.+\..+)$" matches "/ljk.html", client: 172.16.25.3, server: localhost, request: "GET /ljk.html *99 rewritten data: "/wordpress/ljk.html", args: "", client:172.16.25.3, server: localhost, request: "GET /ljk.html *99 "^/wordpress/(.+\..+)$" matches "/wordpress/ljk.html", client: 172.16.25.3, server: localhost, request: "GET /ljk.html *99 rewritten data: "/ljk/ljk.html", args: "", client: 172.16.25.3, server: localhost, request: "GET /ljk.html 浏览器url依然不变 |
以上两个实验对用户透明,用户通过返回码或者地址栏url均不可感知rewrite的存在,可以看做‘内部重定向’;那么什么时候会产生‘外部重定向’呢,现在先试试Permanent、redirect两个标志
三、重写规则后加上 permanent 标志
四、重写规则后加上 redirect 标志
location / { root /usr/local/nginx/html; index index.shtml index.php index.html; if (!-e $request_filename) { rewrite ^/(.+\..+)$ /wordpress/$1 redirect; } access_log: 两个请求 "GET /ljk.html?sds HTTP/1.1" 302 临时重定向302 "GET /wordpress/ljk.html HTTP/1.1" 200 error_log: *108 "^/(.+\..+)$" matches "/ljk.html", client: 172.16.25.3, server: localhost, request: "GET /ljk.html *108 rewritten redirect: "/wordpress/ljk.html?sds", client: 172.16.25.3, server: localhost, request: "GET /ljk.ht ml?sds 浏览器地址栏变成了rewrite后的url |
由三四得出结论:Permanent、redirect两个标志控制是否将重过程在用户端体现出来(即将重写后的url显示在客户端)同时返回301 or 302。
然后再试试last和break两个标志(需要两层及以上跳转来测试)
按照网上较为普遍的说法:
假如一个location里有多条rewrite规则,都是不在该location继续往下匹配,但是
last: 匹配完该条语句后得到的url,重新到server标签下走一遍
break:到此为止(直接以重写后的url在服务器寻找资源)
实验环境:删除html/wordpress下的ljk.html,将ljk.html放置在html/ljk/ljk.html,然后在server标签下配置location /wordpress 的rewrite规则
五、先来看下两次rewrite 规则不加标志的情况
location / { root /usr/local/nginx/html; index index.shtml index.php index.html; if (!-e $request_filename) { rewrite ^/(.+\..+)$ /wordpress/$1; } } location /wordpress { if (!-e $request_filename) { rewrite ^/wordpress/(.+\..+)$ /ljk/$1; } } access_log: GET /ljk.html HTTP/1.1" 200 error_log: 经历两次匹配和重写 *111 "^/(.+\..+)$" matches "/ljk.html", *111 rewritten data: "/wordpress/ljk.html", args: "", *111 "^/wordpress/(.+\..+)$" matches "/wordpress/ljk.html", *111 rewritten data: "/ljk/ljk.html", args: "", 浏览器地址栏url不变 |
六、实验break标志
location / { root /usr/local/nginx/html; index index.shtml index.php index.html; if (!-e $request_filename) { rewrite ^/(.+\..+)$ /wordpress/$1 break; } } location /wordpress { if (!-e $request_filename) { rewrite ^/wordpress/(.+\..+)$ /ljk/$1; } } access_log: GET /ljk.html HTTP/1.1" 404 error_log: *112 "^/(.+\..+)$" matches "/ljk.html", *112 rewritten data: "/wordpress/ljk.html", args: "", *112 open() "/usr/local/nginx/html/wordpress/ljk.html" failed (2: No such file or directory), 加了break,所以在重写成‘wordpress/ljk.html’就没有再走到location /wordpress |
七、实验last标志
location / { root /usr/local/nginx/html; index index.shtml index.php index.html; if (!-e $request_filename) { rewrite ^/(.+\..+)$ /wordpress/$1 last; } } location /wordpress { if (!-e $request_filename) { rewrite ^/wordpress/(.+\..+)$ /ljk/$1; } } 访问http://172.16.25.162/ljk.html access_log: GET /ljk.html HTTP/1.1" 200 error_log: *113 "^/(.+\..+)$" matches "/ljk.html", *113 rewritten data: "/wordpress/ljk.html", args: "", *113 "^/wordpress/(.+\..+)$" matches "/wordpress/ljk.html", *113 rewritten data: "/ljk/ljk.html", args: "", |
由五六七可得出结论:
加break标志时,url一旦找到匹配额规则,就会停止继续匹配并以该rewrite后额url去服务器请求资源;
加last标志或者不加任何标志,其‘过程’和‘结果’一致,会以rewrite后的url再重新到server段下走一遍配置。
并且这两个标志都不会改变浏览器地址栏的url,且返回码亦为200或404等(即对用户透明)
八、涉及到域名重定向的实验
server { listen 80; server_name localhost; rewrite_log on; rewrite ^(.*)$ http://www.baidu.com; } 若rewrite规则后不加标志或者加redircet标志,都会返回“GET / HTTP/1.1" 302”临时重定向 当rewrite规则后加permanent 标志,会返回“GET / HTTP/1.1" 301”永久重定向 |
希望这篇文章能对理解nginx的rewrite有一些帮助!
本文转自kai404 51CTO博客,原文链接:http://blog.51cto.com/kaifly/1706564,如需转载请自行联系原作者