urllib2 timeout issue

[Peter Otten]

Jérôme wrote:

> Hi all.
> 
> I'm having troubles with urllib2 timeout.
> 
> See the following script :
> 
> ----------------------------
> import urllib2
> result = urllib2.urlopen("http://dumdgdfgdgmyurl.com/")
> print result.readline()
> ----------------------------
> 
> If run on my Debian Wheezy computer, or on my Debian Squeeze server,
> the answer is instantaneous :
> 
> [...]
> urllib2.URLError: <urlopen error [Errno -2] Name or service not known>
> 
> When run on my Raspberry Pi with Raspian Wheezy, the answer is
> identical but it takes 10 seconds.
> 
> I tried
> 
>     result = urllib2.urlopen("http://dumdgdfgdgmyurl.com/", timeout=5)
> 
> but I get the same results : instantaneous on Debian, 10 secondes on
> RPi.
> 
> I also added this, as suggested on some StackOverflow pages :
> 
>     import socket
>     socket.setdefaulttimeout(5)
> 
> and it didn't make any difference.
> 
> In both cases, Python version is "Python 2.7.3".
> 
> Am I missing something ?

The problem might be ipv6-related. I'm currently working around a similar 
annoying delay with the monkey-patch demonstrated below:

$ cat patch_getaddrinfo.py
import sys
import time
import urllib2

def monkeypatch_getaddrinfo():
    import socket

    gai = socket.getaddrinfo
    def getaddrinfo(*args):
        return gai(args[0], args[1], 2, *args[3:])

    socket.getaddrinfo = getaddrinfo

if "-4" in sys.argv:
    monkeypatch_getaddrinfo()

start = time.time()
try:
    urllib2.urlopen("http://example.com/")
finally:
    print time.time() - start
$ python -S patch_getaddrinfo.py
20.320786953
$ python -S patch_getaddrinfo.py -4
0.305027008057

As you may infer from the above I don't know what is happening exactly and 
how to fix it properly...