On 11/11/16 03:15, Bill Cole wrote: > > I think it's a perfectly reasonable TTL for a mail system that uses > round-robin DNS for loadbalancing and high availability. Arbitrarily > forcing it higher on the client end could have unintended negative > consequences.
(quick reply) The other question is, what is the TTL for the NS records for the same domain. If the NS has a long TTL, then looking up the A record is pretty cheap (one DNS lookup to authoritative server). If the NS has short TTL, then looking up the A record is pretty expensive (potentially all the way back to the root again). NS records look like short TTL to me. And generally short TTLs all the way, although I've not had a proper dig to see. Tim _______________________________________________ mailop mailing list [email protected] https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
