BIND version (was Re: DNS & BIND (O'Reilly book) )

[Jim Reid]

>>>>> "Jeff" == Jeff Donovan <jdonovan at beth.k12.pa.us> writes:

    Jeff> I have this theory that usually works pretty well. It's when
    Jeff> you have production server, that services many clients It's
    Jeff> usually best not to mess around with the system. We can all
    Jeff> agree with that.  When it comes to "New" software, I take
    Jeff> the same approach. BIND 8 works fine, and it has been
    Jeff> working fine. Just because there is something new does not
    Jeff> make it "Far Superior".

In this case it does. Any objective comparison of the code quality in
BIND8 and BIND9 would reach the conclusion that BIND9 is better. For
instance BIND8's ns_req() and ns_resp() functions are hardly paragons
of good programming practice. BIND9 provides exactly one way of doing
a particular thing. BIND8 often has multiple ways of doing a specific
operation, all slightly inconsistent with each other. This is not good
even if the software "usually works pretty well".

    Jeff> IMHO, I am not ready to run BIND 9 on a machine that I
    Jeff> cannot afford to have down. But I would recommend setting it
    Jeff> up on a secondary to see how well it runs.

In view of sentence preceding it, your last sentence does not make
much sense: a slave (secondary) server is just as important as a
master. Maybe you meant "test server" rather than "secondary"?