Explanation of Primary/Secondary and Mas

[Joseph S D Yao]

On Fri, Aug 25, 2000 at 02:04:36PM +0100, Sam Wilson wrote:
> "Primary" and "secondary" are also used in at least one RFC (and I can't
> remember which) to mean the first server configured to queried by a
> client, the second and so on (the ordered "nameserver" lines in
> /etc/resolv.conf or the equivalent).  I remember this because it was
> pointed out to me after I complained about the usage in a public forum. 
> Oops! :-(
> 
> -- 
> Sam Wilson
> Network Services Division, Computing Services
> The University of Edinburgh
> Edinburgh, Scotland, UK

Quite true; thank you.  This is a different usage, and may be confusing
if you don't pay attention.  These are the primary, secondary, tertiary
etc. servers for the RESOLVER.  Maybe the adjective change for the NAME
SERVERS was to reduce this confusion?  ;-)  I honestly don't know.

The Master name server [for the name servers] is the one from which all
changes eminate.  The Slave name servers look to it for any updates in
their information.  As I mentioned, this does not affect how any
resolvers querying them see them - as a set of peers.  As each slave
declares its own master or masters, and each master may or may not
enumerate its own slaves, and it is possible for slaves to have their
own slaves that regard them as their master(s) - it is possible for the
internal relationships to get quite complicated if you're not careful.
But none of this is seen by the querying name servers unless the
relationships get pathological.  [In which case the BBC might put them
on a show named Fawlty Servers.  Or "Are You Served?" - a whole new
meaning, eh?]

-- 
Joe Yao				jsdy at cospo.osis.gov - Joseph S. D. Yao
COSPO/OSIS Computer Support					EMT-B
-----------------------------------------------------------------------
This message is not an official statement of COSPO policies.