Tony Finch wrote: > Re. "primary", it is worth noting the definition of the SOA MNAME > field in RFC 1035: > > MNAME The <domain-name> of the name server that was the > original or primary source of data for this zone. > > I get the impression that sometimes people regard a zone as having one > primary or master and all its other servers are secondaries or slaves. > But sometimes (eg when configuring slave zones in BIND) you get a > per-xfer view of the terminology.
the reason RFC 1996 taxonomizes as it does is, as said up-thread, that "primary master" means the place where non-dns data enters the dns, and "slave" is the other servers which can also answer authoritatively for them. but, again from RFC 1996, in a multi-layer AXFR hierarchy, some servers will both initiate and respond to AXFR (and then later, IXFR also) requests. therefore the terminology used to describe AXFR/IXFR participants is in a different universe than the terminology used to describe "where data enters the DNS" vs. "where DNS data is also authoritatively served from." if we were allowed to invent new terminology, i'd say: primary, secondary, tertiary, quaternary -- assuming nobody is crazy enough to have more than four layers zone transfer source, zone transfer sink -- instead of "master" and "slave" the first set describes your position in the data path. "primary" is the from-outside-dns wormhole junction. the second set describes your role in an AXFR/IXFR relationship. one of them initiates notify and responds to IXFR/AXFR, the other responds to notify and initiates IXFR/AXFR. (and it would all have to be re-thought for what used to be called "multi-master", anyway.) -- Paul Vixie _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
