SOLVED -- Re: Problems with a BIND server

Barry Margolin barmar at alum.mit.edu
Thu Oct 15 04:14:17 UTC 2009


In article <mailman.704.1255578769.14796.bind-users at lists.isc.org>,
 Robert Moskowitz <rgm at htt-consult.com> wrote:

> When I queried from home.htt (really hda.home.htt), it appears that it 
> does not matter that the SOA and NS are wrong and do not point to an IP 
> address. It is authoratative for the zone and just reports from its 
> cache. Likewise a client that uses it directly as its nameserver, would 
> never be the wiser of the problem. Only when another nameserver did the 
> lookup. If you look at that TCPDUMP use see the first lookup of say, 
> wiki.home.htt which returns the A record. Then a lookup of home.htt 
> which fails. From this point on, ANY lookup of any host in home.htt 
> fails completely. The cache is 'ruined?' with that failed lookup of the 
> NS from hda.home.htt.

When it recurses the first time, the response includes the NS records 
from the authoritative server, as well as the A records if they're 
in-bailiwick.  These take precedence over the delegation and glue 
records in the parent zone, which is why the cache is "ruined".

This is a common cause of intermittent DNS failures out on the public 
Internet, when the NS records in a zone don't match the registered 
nameservers.

-- 
Barry Margolin, barmar at alum.mit.edu
Arlington, MA
*** PLEASE don't copy me on replies, I'll read them in the group ***



More information about the bind-users mailing list