[ Date Index ][
Thread Index ]
[ <= Previous by date /
thread ]
[ Next by date /
thread => ]
Steve Marvell wrote:
Indeed. A good test is to do an nslookup...
nslookup is rarely a good test for anything! Most nslookups will default to using the first name server in resolv.conf, or duck out to 127.0.0.1. Although it tells you what it does (assuming name resolution is working fairly well). host may use resolv.conf as well to establish fully qualified names, although host behaviour varies with version of BIND utilities installed, and resolver in use. My experience of Linux is that it is as easy to check nsswitch.conf, hosts, and NIS seperately if you suspect spurious name resolution data. 99.9% of name service problems I get involved in can be solved with "ping", "traceroute", "more" and "dig" (although I tend to use "doc" to save on typing). nslookup is a dreadful tool for name resolution problems, not least in that it's behaviour varies radically between platforms, and releases. Have a look at comp.protocol.dns.bind archives for various diatribes and explanations as to why "nslookup" is such an appalling diagnostic tool. We dig, dig, dig, dig, dig, dig, dig, we dig the whole day through ;) -- The Mailing List for the Devon & Cornwall LUG Mail majordomo@xxxxxxxxxxxx with "unsubscribe list" in the message body to unsubscribe.