Effects of Very Short TTLs

Royans Tharakan tharakan at lucent.com
Mon Nov 8 21:56:21 UTC 1999


yea... thats true.
another small yet problem which might come up is that, if primary
goes down for more 5 minutes,then slave server might give an
"non-athoritative"
reply, which is not something one is comfortable with.


Barry Margolin wrote:
> 
> In article <806v2l$a26$1 at nntp.msstate.edu>,
> Boyd Nation <boyd at Ra.MsState.Edu> wrote:
> >We have a small DNS setup that requires the ability to make dynamic changes
> >that propagate fairly quickly.  I know, or at least I think I know, that
> >setting very short TTLs, on the lines of five minutes, will cause an
> >increase in network traffic, but that's negligible in this case.  Are there
> >any other effects that I should be aware of?
> 
> Be aware that BIND on the slave servers will not ordinarily check for zones
> whose Refresh times have run out more often than every 15 minutes, so a
> change on the master server could take 15 minutes to propagate to the slave
> servers.  Thus, it could take up to 20 minutes for a change to propagate
> (and even longer if the slaves are very busy, since they might not get a
> chance to refresh the zone immediately when the 15 minutes is up).
> 
> If you use BIND 8 it will make use of the NOTIFY protocol, so the change
> should propagate to slave servers almost immediately.
> 
> --
> Barry Margolin, barmar at bbnplanet.com
> GTE Internetworking, Powered by BBN, Burlington, MA
> *** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
> Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.



More information about the bind-users mailing list