NS rr configuration: 1*NS + 4xA vs. 4xNS ?

Matus UHLAR - fantomas uhlar at fantomas.sk
Tue Jun 30 02:37:27 UTC 2009


> In message <20090629200938.GA6091 at fantomas.sk>, Matus UHLAR - fantomas writes:
> > On 30.06.09 01:08, Mark Andrews wrote:
> > > In message <20090629101834.GA31308 at fantomas.sk>, Matus UHLAR - fantomas wri
> > tes:
> > > > I am planning to change NS records in our and our customers' zones.
> > > > 
> > > > I'll have four nameservers on different networks, and I'd like to make
> > > > configuration as easy as possible by using only one NS record for them al
> > l.
> > > 
> > > 	And harder to debug.  1 name to 1 machine is easy to debug.
> > 
> > running either of them behind a L3 switch makes it hard to debug again,
> > so I wouldn't take that as an issue.

On 30.06.09 10:01, Mark Andrews wrote:
> 	A L3 switch is still one virtual machine with one routing
> 	entry and one path from the customer to the L3 switch.

the difficulty of debugging the case when one IP directs to L3 switch with
more real machines behind is bigger than the difficvulty of debugging case
where one A points to more real IPs...

> 	There is no need to play this silly game.  It just make
> 	things harder.  Some machines will make assumptions that
> 	all the address refer to one machine and that some operations
> 	shouldn't be retried because they won't get a different
> 	response.

This is just what I wanted to know. Is there any evidenve that any DNS
resolvers have such behaviour?

-- 
Matus UHLAR - fantomas, uhlar at fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
Silvester Stallone: Father of the RISC concept.



More information about the bind-users mailing list