organization of the file after ddns

[Jim Reid]

>>>>> "Alexandre" == Alexandre Vilarinho <avilarinho at multirede.com.br> writes:

    Alexandre> After a ddns update bind 9 and 8 re-arrange the
    Alexandre> file... bind 9 do it in an = alphabetical order (am i
    Alexandre> right?). I didn't see a pattern used on 8, does = any
    Alexandre> one saw it, and could explain it to me please?

I curious why you care about this. How a name server manages the zone
files it controls is nobody's business but that name server's. DNS
administrators shouldn't need to look at them or make assumptions
about the format and any ordering that's used. This point was made
earlier on this thread.

FWIW, BIND9 stores each zone in its own data structure: a red-black
tree. When this tree is traversed, the nodes are in essentially
alphabetical order. So that's what you get when the name server writes
out the zone file. BIND8 has a monolithic data structure for all the
DNS data. It's a glorified hash table. So zone traversal generally
means the nodes (names) are found in a random order. Of course this
could change if either implementation chooses to have a different
internal data structure. [This might even be seen now with say an LDAP
back-end for BIND9.] Since changing those data structures would be an
implementation issue for the BIND developers, this means nobody should
be making assumptions about the ordering that zone files get written
out. Or creating scripts and tools that depend on a particular format
or RR ordering.