cURL 冻结 Apache 并在出现 404 错误时超时

Question

我遇到一个关于 cURL 请求的非常奇怪的问题。基本上，我有一些代码用于在 Symfony2 (http://symfonybricks.com/it/brick/wrap-legacy-php-code-by-a-symfony2-application) 上包装遗留 PHP 应用程序。它们是基本的 cURL 命令，应该拉取网页并显示它。但是，我注意到以下内容：

如果我对现有文件进行查询，它工作正常
如果我对不存在的文件进行查询，cURL 超时（因为我设置了 8 秒的超时），在此期间，它冻结了整个服务器，我不能从任何其他设备访问我的网站，直到它真正超时。

这是我的代码：

/**
 * @Route("/", defaults={"controller" = "index.php", "controller2" = "", "controller3" = ""})
 * @Route("/{controller}", defaults={"controller2" = "index.php", "controller3" = ""})
 * @Route("/{controller}/", defaults={"controller2" = "index.php", "controller3" = ""})
 * @Route("/{controller}/{controller2}", defaults={"controller3" = ""})
 * @Route("/{controller}/{controller2}/", defaults={"controller3" = ""})
 * @Route("/{controller}/{controller2}/{controller3}", defaults={"controller3" = "index.php"})
 */
public function getLegacyResourceAction($controller, $controller2, $controller3, Request $request)

{


    $path_to_legacy_code = "http://XXX/";

    $originalController = $request->getPathInfo();

    $originalQueryString = $request->getQueryString();

   $url = "{$path_to_legacy_code}{$originalController}?{$originalQueryString}";

    //open connection

    $ch = curl_init();

    curl_setopt($ch, CURLOPT_URL, $url);

    curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);

    curl_setopt($ch, CURLOPT_FOLLOWLOCATION, 1);

    curl_setopt($ch, CURLOPT_HEADER, false);

    curl_setopt($ch, CURLOPT_VERBOSE, 1);

    curl_setopt($ch, CURLOPT_USERAGENT, $_SERVER['HTTP_USER_AGENT']);

    //Timeout
    curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 8);
    curl_setopt($ch, CURLOPT_TIMEOUT, 8); 
    curl_setopt($ch,  CURLOPT_MAXREDIRS, 5);       

    $stderr = fopen("{$this->container->getParameter('kernel.root_dir')}/logs/curl.txt", "a");
    curl_setopt($ch, CURLOPT_STDERR, $stderr);


    //echo "Login stuff in ".$this->container->getParameter('kernel.root_dir')."/logs/curl.txt";


    curl_setopt($ch, CURLOPT_COOKIESESSION, 0);

    curl_setopt($ch, CURLOPT_COOKIEFILE, 'cookies.txt');

    curl_setopt($ch, CURLOPT_COOKIEJAR, 'cookies.txt');  

    $result = curl_exec($ch);


    if (false === $result) {
        echo curl_error($ch);
        exit;
    }

    curl_close($ch);
    fclose($stderr);

    $Response = new Response($result);


    return $Response;

这是我从有效响应中获得的日志：

* About to connect() to www.xxx.com port 80 (#0)
*   Trying 178.32.223.113...
* connected
* Connected to www.xxx.com (178.32.223.113) port 80 (#0)
> GET /web/legacy/index.php HTTP/1.1
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:40.0) Gecko/20100101 Firefo$
Host: www.xxx.com
Accept: */*
Cookie: idto=116; PHPSESSID=a159rcjvh0fk6otukqqq9bkrd5

* additional stuff not fine transfer.c:1037: 0 0
* HTTP 1.1 or later with persistent connection, pipelining supported
< HTTP/1.1 200 OK
< Date: Sat, 05 Sep 2015 04:26:32 GMT
< Server: Apache/2.2.22 (Debian)
< X-Powered-By: PHP/5.5.28-1~dotdeb+7.1
< Expires: Thu, 19 Nov 1981 08:52:00 GMT
< Cache-Control: no-store, no-cache, must-revalidate, post-check=0, pre-check=0
< Pragma: no-cache
< Vary: Accept-Encoding
< Transfer-Encoding: chunked
< Content-Type: text/html
<
* Connection #0 to host www.xxx.com left intact
* Closing connection #0

这是来自无效响应的代码：

* About to connect() to www.xxx.com port 80 (#0)
*   Trying 178.32.223.113...
* connected
* Connected to www.xxx.com (178.32.223.113) port 80 (#0)
> GET /web/legacy/whatever.php HTTP/1.1
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:40.0) Gecko/20100101 Firefo$
Host: www.xxx.com
Accept: */*
Cookie: idto=116; PHPSESSID=a159rcjvh0fk6otukqqq9bkrd5

* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* Operation timed out after 8001 milliseconds with 0 bytes received
* Closing connection #0
* About to connect() to www.xxx.com port 80 (#0)
*   Trying 178.32.223.113...
* connected
* Connected to www.xxx.com (178.32.223.113) port 80 (#0)
> GET /web/legacy/legacy/whatever.php HTTP/1.1
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:40.0) Gecko/20100101 Firefo$
Host: www.xxx.com
Accept: */*
Cookie: idto=116; PHPSESSID=a159rcjvh0fk6otukqqq9bkrd5

* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* additional stuff not fine transfer.c:1037: 0 0
* Operation timed out after 8001 milliseconds with 0 bytes received
* Closing connection #0
* About to connect() to www.xxx.com port 80 (#0)
*   Trying 178.32.223.113...
* connected
* Connected to www.xxx.com (178.32.223.113) port 80 (#0)
> GET /web/legacy/legacy/legacy/whatever.php HTTP/1.1
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:40.0) Gecko/20100101 Firefo$
Host: www.xxx.com
Accept: */*

下面的一切curl_exec($ch);直到超时才执行，没有给我机会检查 404 错误或其他东西。我已经找了好几天了，到目前为止还没有找到。

非常感谢！

编辑：

好的，所以我通过删除这一行解决了这个问题：

curl_setopt($ch, CURLOPT_COOKIEFILE, 'cookies.txt');

现在它给出了正确的 404 错误。我不知道为什么没有这条线它就可以工作。问题是，如果没有这一行，我将破坏遗留代码中的一些会话内容！

编辑 2：

我已经改变了：

curl_setopt($ch, CURLOPT_COOKIESESSION, 0);

到

curl_setopt($ch, CURLOPT_COOKIESESSION, 1);

现在它在 404 错误和我的会话中工作正常。我不知道为什么。

编辑 3：

好吧，我想我终于明白发生了什么事了：

在每个 cURL 请求中，它写入 cookie.txt 文件
如果有人查询 404 url，在超时之前，它会锁定 cookie.txt 文件，这就是为什么整个事情似乎都被冻结了。

据我所知，最好的解决方案是为每个用户生成一个不同的 cookie.txt 文件，以防止文件锁定。

Answer 1

我终于弄清楚了，原来我遇到了两个问题：

没有配置超时，所以如果我尝试查询一个不存在的网页，它将无限循环。
我为每个人使用相同的 cookie 文件，如果有人陷入前面提到的这种循环，它会文件锁定 cookie 文件，阻止其他用户访问该站点。

cURL 冻结 Apache 并在出现 404 错误时超时

cURL freeze Apache and time out on a 404 error

php

curl

timeout

libcurl

symfony