CNAME Error

[Jim Reid]

>>>>> "Jonathan" == Jonathan de Boyne Pollard <J.deBoynePollard at Tesco.NET> writes:

    JR> The OP wouldn't have had this problem if he/she explicitly
    JR> used fully qualified, dot-terminated domain names throughout
    JR> the zone file.

    Jonathan> Yann wouldn't have had this problem if he/she had
    Jonathan> *explicitly filled in all of the fields* in each
    Jonathan> resource record, especially the owner name.  Whether the
    Jonathan> owner name was fully-qualified is in fact irrelevant,
    Jonathan> and indeed using non-fully-qualified domain names is a
    Jonathan> useful practice, especially when one wants to do things
    Jonathan> like migrate to different domain names or compile
    Jonathan> multiple "zones" from the same single source file.  The
    Jonathan> important factor here was the omission of the fields,
    Jonathan> not whether those fields comprised fully-qualified
    Jonathan> domain names.

Putting fully-qualified, dot-terminated domain names in zone files --
especially in owner-name fields -- is part and parcel of "explicitly
filling in all the fields". It's an example of saying what you mean
and meaning what you say. It also prevents easily avoidable mistakes.
For instance when a $ORIGIN directive gets added to the zone file in
the wrong place and all of a sudden the names of everything beneath it
in the zone file change.

Saving a few bytes of disk space (how much is it for a Gbyte these
days?) or a couple of microseconds of zone load time because of FQDNs
in the zone file is a false economy. The hardware resources "wasted"
in this way are orders of magnitude cheaper than even an hour of a DNS
administrator's time. Which is orders of magnitude cheaper than the
costs of downtime to an organisation's email or web service (say)
because of a weird DNS problem.