Controversial SOA values ?

wwebb at adni.net wwebb at adni.net
Fri Dec 24 11:58:30 UTC 1999


>> From:             	<wwebb at adni.net>
> > In addition to the 1000 hour expire suggestion I mentioned
> > previously at: http://www.ripe.net/docs/ripe-203.html , I note that
> > RFC 1912, Section 2.2, says "...2-4 weeks are suggested (expire)
> > values."   My concern, expressed another way is, if I'm on vacation
> > in Aruba and they call and say the server blew-up, I don't have to
> > quit my vacation to go make the DNS slaves happy.   :-)

> From:			<milli at acmebw.com>
> Sounds perfectly reasonable.  May work fine for your situation.  I guess
> I've just never worked in a shop where a week without zone update capability
> was anywhere near acceptable.  Always had a complete spare server sitting
> around at another geographically-separated location in case of a meltdown.
> If the Bay Area fell off the map, we could switch over to an East Coast
> server in the same day.  Course, if the United States fell off the map,
> well, we've likely got bigger problems to worry about.  ;-)

In checking around a bit,  I noticed that the DNS folks at 
organizations such as HARVARD.EDU,  BERKLEY.EDU, 
BMW.COM, APPLE.COM, MICROSOFT.COM, CONGRESS.GOV 
all use expire times in excess of 3 weeks (and some over 11 
weeks).     

Bill Webb




More information about the bind-users mailing list