记得以前写爬虫的时候为了防止dns多次查询，是直接修改/etc/hosts文件的，最近看到一个优美的解决方案，修改后记录如下：

import socket

_dnscache={}
def _setDNSCache():
    """
    Makes a cached version of socket._getaddrinfo to avoid subsequent DNS requests.
    """

    def _getaddrinfo(*args, **kwargs):
        global _dnscache
        if args in _dnscache:
            print str(args)+" in cache"
            return _dnscache[args]

        else:
            print str(args)+" not in cache"  
            _dnscache[args] = socket._getaddrinfo(*args, **kwargs)
            return _dnscache[args]

    if not hasattr(socket, ‘_getaddrinfo‘):
        socket._getaddrinfo = socket.getaddrinfo
        socket.getaddrinfo = _getaddrinfo

def test():
	_setDNSCache()
	import urllib
	urllib.urlopen(‘http://www.baidu.com‘)
	urllib.urlopen(‘http://www.baidu.com‘)

test()

结果如下：

(‘www.baidu.com‘, 80, 0, 1) not in cache
(‘www.baidu.com‘, 80, 0, 1) in cache

1.相当于只对socket.getaddrinfo打了一个patch，但socket.gethostbyname,socket.gethostbyname_ex还是走之前的策略

2.只对本程序有效，而修改/etc/hosts将对所有程序有效，包括ping