MDN’s new design is in Beta! A sneak peek:


URL 重定向,或者说 URL 转递,是一种 当实际资源如一个网页或一个表单或一整个weby应用被定位为到一个不同的URL地址时,保持链接可用的技术。 HTTP提供了一种专门的响应——HTTP 重定向,来为各种各样的目标执行以下操作:当网站处于维护时的临时重定向,网站架构改变后为了保持外部链接继续可用的永久重定向, 当上传一个文件时在页面显示进度条,以及更多……


在HTTP, 重定向由服务器发送一个特定的响应: redirects触发. HTTP重定向响应具有状态码 3xx. 一个浏览器, 当接收到重定向响应,使用新提供的URL并立即加载它: 大部分时间, 除了一小部分性能损失,重定向对用户透明,.

重定向有好几种类型他们分属三个类别:永久, 临时和特殊重定向.


这些重定向意味着永久. 它表示原URL不应再被使用新的URL是首选. 搜索引擎在其索引资源触发相关URL的更新.

Code Text 处理方法 典型应用
301 Moved Permanently GET 方法不变其他可能会可能不会变成GET.[1] 网站重组.
308 Permanent Redirect 方法和主体(body)不变. Reorganization of a Web site, with non-GET links/operations.

[1] 该规范无意使方法发生改变, 但实际上用户代理会这么做( but practically there are user agents out there doing this). 308被创建用来消除使用非GET方法时的歧义行为.


有时请求的资源不能从它的规范位置访问, 但能从另外的地方访问. 在这种情况下可以使用一个临时重定向. 搜索引擎不会记录新的地址.,即该临时链接 .当创建,更新和删除资源时临时重定向也被用作以呈现临时的进度页面.

Code Text 处理方法 典型应用
302 Found GET 方法不变其他可能会可能不会变化GET.[2] 由于不可预见的原因该页面暂不可用. 这种情况, 搜索引擎不会更新他们的链接.
303 See Other GET 方法不变,其他方法变为GET (body lost). Used to redirect after a PUT or a POST to prevent a refresh of the page that would re-trigger the operation.
307 Temporary Redirect 方法和主体不变 由于不可预见的原因该页面暂不可用. 这种情况, 搜索引擎不会更新他们的链接. Better than 302 when non-GET links/operations are on available on the site.

[2] The specification had no intent to allow method changes, but practically there are user agents out there doing this. 307 has been created to remove the ambiguity of the behavior when using non-GET methods.

Special redirections

In addition to these usual redirections, there are two specific redirections. The 304 (Not Modified) redirects a page to the locally cached copy (that was stale), and 300 (Multiple Choice) is a manual redirection: the body, presented by the browser as a Web page, lists the possible redirections and the user clicks on one to select it.

Code Text Typical use case
300 Multiple Choice Not many: the choices are listed in an HTML page in the body. Could be served with a 200 OK status.
304 Not Modified Cache refresh: this indicates that the cache value is still fresh an can be used.


HTTP重定向并非定义重定向的唯一方式。定义重定向有两种方法:1.使用 HTML 的 meta元素进行 HTML 重定向。2.使用 DOM 进行 JavaScript 重定向。

HTML redirections

HTTP redirects aren't the preferred way to create redirections, but sometimes the Web developer doesn't have the control over the server or cannot configure it. For these specific cases, the Web developers can craft an HTML page with a <meta> element and the http-equiv attribute set to refresh in the <head> of the page. When displaying the page, the browser will find this element and will go to the indicated page.

  <meta http-equiv="refresh" content="0;URL=''" />

The content attribute starts with a number indicating how many seconds the browser should wait before redirecting to the given URL. Always set it to 0, for better accessibility.

Obviously, this method only works with HTML pages (or similar) and cannot be used for images or any other type of content.

Note that these redirections break the back button in browser: you can go back to a page with this header but it instantaneously moves forward again.

JavaScript redirections

Redirections in JavaScript are created by setting a value to the window.location property and the new page is loaded.

window.location = "";

Like HTML redirections, this can't work on all resources, and obviously this will only work on clients that execute JavaScript. On the other side, there are more possibilities as you can trigger the redirection only if some conditions are met, for example.

Order of precedence

With three possibilities for URL redirections, several methods can be specified at the same time, but which one is applied first? The order of precedence is the following:

  1. HTTP redirects are always executed first when there is not even a page transmitted, and of course not even read.
  2. HTML redirects (<meta>) are executed if there weren't any HTTP redirects.
  3. JavaScript redirects are used as the last resort, and only if JavaScript is enabled on the client side.

When possible, always try to use HTTP redirects, and don't use a <meta> element. If a developer changes the HTTP redirects and forgets the HTML redirects, redirects are no more identical, or end up in an infinite loop, or other nightmares happen.


There are numerous use cases for redirects, but as performance is impacted with every redirect, their use should be kept at a minimum.

Domain aliasing

Ideally, there is one location, and therefore one URL, for one resource. But there are reasons for wanting to have alternative names for a resource (several domains, like with and without the www prefix or shorter and easy to remember URLs, …). In these cases, rather than duplicating the resource, it is useful to use a redirect to the one true (canonical) URL.

Domain aliasing can be done for several reasons:

  • Expanding the reach of your site. A common case is when your site resides under the domain and accessing your pages from should be possible, too. Redirections for pages to are set up in this case. You might also provide commonly used synonym names, or frequent typos of your domain names.
  • Moving to a different domain. For example, your company has been renamed and you when searching for the old name, you want people used to the old company Web site still find you under the new name.
  • Forcing HTTPS. Request to the HTTP version of your site will be redirected to the HTTPS version of your site.

When you restructure Web sites, URLs of resources change. Even if you can update the internal links of your Web site to match the new naming scheme, you have no control over the URLs used by external resources. You don't want to break this link as they bring you valuable users (and help your SEO), so you set up redirects from the old URLs to the new ones.

Even if this technique also works for internal links, you should try to avoid having internal redirects. A redirect has a significant performance cost (as an extra HTTP request is done) and if you can avoid it by correcting internal links, you should fix these links.

Temporary responses to unsafe requests

Unsafe requests modify the state of the server and the user shouldn't to replay them inadvertently. Typically, you don't want your users to resent PUT, POST or DELETE requests. If you just serve the response as the result of this request, a simple press of the reload button will (possibly after a confirmation message), resend the request.

In this case, the server can send back a 303 (See Other) request that will contain the right information, but if the reload button is pressed, only this page is redisplayed, without replaying the unsafe requests.

Temporary responses for long requests

Some requests may need more time on the server, like sometimes DELETE requests that are scheduled for later processing. In this of case, the response is a 303 (See Other) redirect that links to a page indicating that the action has been scheduled, and eventually informs about the progress, or allows to cancel it.



Redirects can be set either in the server config file, either in the .htaccess of each directory.

The mod_alias module has Redirect and Redirect_Match directives that set up a 302 response (by default):

<VirtualHost *:80>
	Redirect /

The URL will be redirected to (but not )

Redirect_Match does the same but takes a regular expression to define a collection of URLs that are affected:

RedirectMatch ^/images/(.*)$$1

All documents in the images/ folder will be redirected to a different domain.

If you don't want to set up a temporary redirect, an extra parameter (either the HTTP status code to use or the permanent keyword can be used to set up a different redirect:

Redirect permanent /
Redirect 301 /

The mod_rewrite module can also be used to create redirects. It is more flexible, but a bit more complex to use.


In Nginx, you create a specific server block for the content you want to redirect:

server {
	listen 80;
	return 301 $scheme://$request_uri;

For having a redirect applying to a folder or a subset of the pages only, use the rewrite directive:

rewrite ^/images/(.*)$$1 redirect;
rewrite ^/images/(.*)$$1 permanent;


In IIS, you use the <httpRedirect> element to configure redirections.


Redirection loops happen when successive redirections follow the one that has already been followed. In other words, there is a loop that will never be finished and no page will be found ultimately.

Most of the time this is a server problem and if the server cannot detect it, it will send back a 500 Internal Server Error. If you encounter such an error soon after modifying a server configuration, this is likely a redirection loop.

Sometimes, the server won't detect it: a redirection loop can spread over several servers which each don't have the full picture. In this case, browsers will detect it and  post an error message. Firefox will display:

Firefox has detected that the server is redirecting the request for this address in a way that will never complete.

while Chrome will display:

This Webpage has a redirect loop

In both cases, the user can't do much (unless a corruption is happening on their side, like a mismatch of cache or cookies).

It is important to avoid redirection loops as they completely break the user experience.


 此页面的贡献者: mushang11, zhi.lin, ZhongyiChen
 最后编辑者: mushang11,