You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
declare(suffix, "info-producer-members-query", "", "select domains.id, domains.name, domains.options from records join domains on records.domain_id=domains.id and records.name=domains.name where domains.type='MASTER' and domains.catalog=$1 and records.type='SOA' and records.disabled=false");
For the typical case---sending a list of candidate secondaries to a secondary auth---that is fine.
In the case of using a catalog zone to send to a recursor (for defining forwards) that's too restrictive. A NATIVE zone could be perfectly fine for that. Maybe even a secondary zone could be used.
We should reconsider this strict approach, IMO.
The text was updated successfully, but these errors were encountered:
I agree, every type of zone could be a catalog member*. Especially with secondary-do-renotify it makes sense to allow secondary-zones beeing part of a catalog. Type native allows outoing AXFR but does not sent NOTIFYs, so it may sound a bit strange but I am sure someone will find a use case too.
I wonder if catalog zones could/should be recursive and type catalog is a valid type too. For example you have a catalog per customer, and a meta-catalog with all the customer-catalogs. So if you have a new customer, you create a catalog for that customer and add that customer-catalog zone to the meta-catalog zone. That way, catalogs may be partitioned and smaller which may improve speed for catalog creation, AXFR and consuming.
Uh oh!
There was an error while loading. Please reload this page.
Currently the SQL enumerating the members of a catalog zone requires them to be of type 'MASTER'. e.g.
pdns/modules/gpgsqlbackend/gpgsqlbackend.cc
Line 143 in 965ba47
For the typical case---sending a list of candidate secondaries to a secondary auth---that is fine.
In the case of using a catalog zone to send to a recursor (for defining forwards) that's too restrictive. A NATIVE zone could be perfectly fine for that. Maybe even a secondary zone could be used.
We should reconsider this strict approach, IMO.
The text was updated successfully, but these errors were encountered: