On Fri, Aug 21, 2009 at 7:53 PM, Aaron Hopkins <lists at die.net> wrote: On Fri, 21 Aug 2009, Paul Wouters wrote: > > Perhaps Wouter can explain that part, as I am sure some conscious design >> decision has gone into that. >> > > I'm guessing this is the same anti-feature-creep sentiment as why > round-robinning RRs was left out of NSD. This is unfortunate, because very > few clients bother to use anything but the first IP returned by their > resolver. > > But in 300 seconds, things will change. For me, the list got returned >> the second time as: >> > > This would not be true if cnn.com were served by NSD. The ordering would > by > the same, every time, resulting in at least 3x the load reaching the first > IP in the zone file. Hi Aaron and Paul, Thanks for the replies. If I read this correctly, this seems to negate the whole point of using dns round robin. Depending on what my TTL is set to, all queries will be delivered to the first host, whether there are 3 or 30 hosts, and will continue to be delivered to that host until the TTL expires and the cache goes and fetches the records again. Hopefully I am reading this incorrectly and it is possible to get unbound to cycle through its records. Cheers, Gareth -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://unbound.nlnetlabs.nl/pipermail/unbound-users/attachments/20090826/ff82fa40/attachment.htm>