Why could Curl be slower than a web browser?

Question

I am using this class to make one GET and another POST request to a website (the first request is to set a cookie). I am testing in a Win XP virtual machine with virtualbox, using wamp from wampserver dot com. The 2 requests takes from 10 to 18 seconds (with curl), but if I make those request directly via the webbrowser in that same virtual machine the website loads in just a few seconds, and it retrieves all the images, css, etc.

What could be causing curl to work so slow? is there a way to fix it?

Answer 1

Curl is probably trying to Reverse DNS the server and, as it cant, it just hangs there a little waiting for the timeout.

If the timeout is given by IPV6 you can try CURL_IPRESOLVE_V4 to bypass it altogether. It really depends on your machine configuration and is more a question to Server Fault.

Answer 2

检查您的Web服务器日志，并尝试查找来自普通Web浏览器的请求和curl请求之间的任何差异

Answer 3

I faced the same issue, using the curl command.

as suggested above - forcing ipv4 only dns lookup fixed it.

curl -4 $url # nice and fast

(I already had ::1 localhost in my hosts file - but that didn't help).

Answer 4

It is probably due to IPv6.

Try add

curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4 );

Answer 5

I ran into this issue with a local web server. I was able to fix by adding

::1 localhost

to /etc/hosts/ file.

This is the ipv6 notation for 127.0.0.1