This is a discussion on RE: Bind behaves weirdly within the Bind Users forums, part of the DNS and Related Forums category; On Wed, 2004-12-29 at 16:39, jc pinoteau wrote: > > I have occasionally seen this. I could ...
|
|||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
|
|||
|
On Wed, 2004-12-29 at 16:39, jc pinoteau wrote:
> > I have occasionally seen this. I could be mistaken here. My > > impression is that it occurs when all nameservers for the domain are > > inside that domain, and apparently the glue A records for those > > nameservers expired before the NS records themselves. It seems that, > > in those conditions, bind won't be able to resolve the address > > properly until the NS records time out. > > It's an interesting hypothesis. In my case it happens almost every day and > mostly with mainstream website like google or yahoo. My bind can't resolve > while the forwarder can. > > Perhaps because the forwarder allows tcp connections whereas your system only allows udp. > > Some time ago, I set "max-cache-ttl 43200 ;" in my options. I don't > > recall seeing the problem since then -- at least not on my servers. > > I still see it on some other campus servers which lack that option > > setting. Incidently, the reason for the option setting was unrelated > > to this problem. -- G. Roderick Singleton <gerry@pathtech.org> PATH tech |