On Wed, 2011-01-12 at 13:34 -0500, William L. Thomson Jr. wrote: > On Wed, 2011-01-12 at 13:22 -0500, Tom Allen wrote: > > > Additionally, many servers will cache your non-resolution for an > > hour or more, and so if your server is "down" for 5 minutes, your > > outage becomes much longer with a cached no-resolve result. > > Now that I could see having a greater impact. I haven't really > experimented with what a DNS server does if it does not get a result. > But caching nothing seems kinda like a bad idea. One would think the > software would know it did not get a result and maybe try again. I would > think that to be different than getting a result and retaining that as > part of cache. > > I might see about playing around with that at some point. If DNS servers > do cache no result just the same as having a result. Then it makes much > more sense as to why having two or more.
The more I think about this scenario I think its highly unlikely short of caching only name servers. Which tend to be different than actual name servers. Mainly for the fact that the period of time a domain name record is kept in cache is specified by the DNS server responsible for the domain. That is you control the time your IP address will be cached after query. If the server is caching that, its because of a non-standard feature or its a caching name server. A normal record if no result is returned, there will be no TTL specified. Therefore I cannot see how a DNS server would cache, for a period of time, that there was no result. Just the same if your worried about propagation, before you go to make any changes to your DNS server. You can adjust the time a record is kept and that should help expedite any propagation. Which depending on where your making the change, and who the registrar is, can happen rather quickly. -- William L. Thomson Jr. Obsidian-Studios, Inc. http://www.obsidian-studios.com --------------------------------------------------------------------- Archive http://marc.info/?l=jaxlug-list&r=1&w=2 RSS Feed http://www.mail-archive.com/[email protected]/maillist.xml Unsubscribe [email protected]
