I've been thinking about this as well, but haven't gotten to a postable solution yet.
Incidently, I use pgsql rather than mysql, and that probably affects my design choices....) My thought is that there should be a table for each major type of RR (such as rr.a, rr.mx, rr.aaaa, etc) with rr-epcific schema, an rr.misc table with a schema close to that of the current public.records table, and an rr.secondary table which pdns can write to when using axfr replication. The SOA details would be in extra columns in the public.domains table. Then, a public.records view would be created using a UNION query to combine the details from each of the rr.* tables, plus append the SOA details from the domains table into SOA rows in the view. Each table in the rr. space will have a trigger to update the expire column in the domains table when a row is inserted or updated. The rr.secondary table's trigger would update all of the SOA details in the public.domains table on insert/update, to ensure everything is in sync with the master. I beleive all of this can be done as well in modern mysql, and know it also can be done in sqlite, although I don't know whether they support views; the queries may instead need to be stored in the pdns conf for those backends, rather than using a view to duplicate the look-n-feel that the default queries expect. Either way, of course, the query used when storing a seconday will need to be customised. Overall this should improve reliability of the data store and improve the UIs used for storing or manipulating the data. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 _______________________________________________ Pdns-users mailing list [email protected] http://mailman.powerdns.com/mailman/listinfo/pdns-users
