How dhcpd chooses which server to send DDNS updates

John Hascall john at
Thu May 17 18:41:29 UTC 2007

I'm wondering about the following comment in
minires/res_findzonecut.c :

 *      ultimately we want some server addresses, which are ideally the ones
 *      pertaining to the SOA.MNAME, but only if there is a matching NS RR.

which describes the behaviour of the res_findzonecut() function, and,
frankly, that just seems wrong to me.  It also does not seem to be the
behavior of, say, the nsupdate program from ISC's bind distribution.

This makes it so a stock dhcpd, (I've modified ours), can not update
a "hidden primary master", such as:     IN SOA (
                                2007051509 ; serial
                                36000      ; refresh (10 hours)
                                3600       ; retry (1 hour)
                                604800     ; expire (1 week)
                                360000     ; minimum (4 days 4 hours)

where xx-0 is the master and dns-1/dns-2 are the externally visible
slave servers.

Is there any chance this is might change in future versions or am I
going to be stuck making this mod from now on?

PS, it's about a 25 line change if anybody else is interested

More information about the dhcp-users mailing list