nsd users - Feb 2006 - NSD 2.3.3 Load Balancing

Can nsd 2.3.3 make load balancing? 

Example: 

www 600 IN A 10.0.0.1
         600 IN A 10.0.0.2
         600 IN A 10.0.0.3 

Greet 

Sebastian Schikora

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Sebastian wrote:
> Can nsd 2.3.3 make load balancing? 
> 
> Example: 
> 
> www 600 IN A 10.0.0.1
>          600 IN A 10.0.0.2
>          600 IN A 10.0.0.3 
Hi Sebastian,

No, nsd does not load balance for you.
Round robin sequence of RRset members in consecutive answers
is an explicit NON-requirement for NSD. (from the REQUIREMENTS file).

Best regards,
   Wouter
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.1 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFD/t0QkDLqNwOhpPgRAmoCAKCctrbqinBWDvlIN9kYHOulewey7wCdGPWa
zhxtK2MWogS+0yKd0HgmroE=somf
-----END PGP SIGNATURE-----

[Quoting "Sebastian", on Feb 24, 10:43, in "NSD 2.3.3 Load Balan
..."]
> Can nsd 2.3.3 make load balancing? 
> 
> Example: 
> 
> www 600 IN A 10.0.0.1
>          600 IN A 10.0.0.2
>          600 IN A 10.0.0.3 
What you want for this is random or sequentional reordering of the
recordset on consecutive queries so that resolvers use the different
IP-numbers to spread the load.  Most (almost all end-user) resolvers
do not do the recursing themselves, but use caching-forwarders.
Stub-resolvers, which use caching-forwarders, will thus never query
an NSD nameserver, as NSD is authoritative-only.

Caching-forwarders may or may not be setup to reorder.

Now, to answer you question more directly:
 NSD does not re-order the recordset, but for load balancing
 this is of little importance, as it is the behaviour of the
 caching-forwarders which does the trick.

-- ted