At 5/5/02 9:02 AM, Lynn W. Taylor wrote: >Not having a change-all-nameservers is also dangerous: > >We trade secondary name service with another ISP. They made some changes >and suddenly, one of our secondary name servers was gone.
If they had kept the host name the same, you wouldn't need to update every nameserver entry -- you (or they) would just update the IP address of the host record. It sounds like they changed (or discontinued) the name of the server with one hour's notice, which is, ummm, less than optimal. >Our customers felt the impact right away, and we knew about the problem an >hour or two before the root-zone update (at noon and midnight Eastern time) >but there was simply no way to change all of them, one at a time, before >the cutoff, so quite a few of them didn't get changed in time. Even if you had updated them all, it wouldn't have helped -- the old values would have been cached for up to two days anyway. For example, the old values were probably cached for all AOL users. A bulk-nameserver update feature is definitely not the solution to the particular problem you experienced. If anyone is relying on getting some change into a root server update on the theory that it will immediately take effect (by the way, they're theoretically at 5 AM and 5 PM now, not noon and midnight), that person will be sadly disappointed. Many people across the Internet will not see the change for a day or two anyway, so worrying about whether you can enter it an hour or two earlier is pointless. I'm not saying that such a feature might not be useful as a labor-saving device; just that it can't be justified on the grounds that someone might need to do an emergency update sometime. Anyway, if it was only one of your two nameservers that was offline, that's not the end of the world as long as the other was working; that's why you use two. -- Robert L Mathews, Tiger Technologies "The trouble with doing something right the first time is that nobody appreciates how difficult it was."
