Greg A. Woods put forth on 3/6/2010 2:58 PM: > At Sat, 06 Mar 2010 14:42:13 -0600, Stan Hoeppner <[email protected]> > wrote: > Subject: Re: reverse dns fails with multiple domains >> >> RFC does not dictate that your forward and reverse dns names match. > > Common sense and common decency do though -- since if the forward and > reverse names are not all orthogonal then the DNS lies, either by > omission, or outright.
Apparently you've missed past discussions here showing some examples of why this can be neither practical or desirable in some situations. > For every hostname pointing at an IP address, there should be a > corresponding PTR for that address pointing back at the hostname. When you say hostname, are you talking A record? Are you talking all IPs in general, or only MX hosts, or SMTP sending hosts? Does a web server ever need a PTR? Do any web browsers ever look up a host via PTR? No. So why should a web server have a PTR? > There's no real excuse for mis-matched forward and reverse DNS. If > you're going to show your reverse DNS to the world, then do it right. A web server with a single IP address hosting 378 vitural domains. Should it have 379 PTRs? One for the host itself and one for each virtual domain? Of course not. A mail server with a single IP address hosting 378 mail domains? Should it have 379 PTRs? One for the host itself and one for each virtual MX domain? Of course not. In this case, the DNS infrastructure isn't smart enough to return matching records even though they do exist, so why bother? You're living in a "perfect" world where everything has a 1:1 relationship in DNS. In the real world, this isn't the case, and probably never will be. I argued your position for years until I was blue in the face. You know what it gained me? A blue face. Nothing else. BTW, please keep list correspondence on list. I don't see any reason why your reply needed to be off list. -- Stan
