I have a setup of several frontend (MX) postfix servers using relay_domains. Some of the relay_domains are hosted on a postfix backend server that has the mailboxes. The other relay_domains are of customers running their own mailservers.
transport_maps is used to retrieve the nexthop destination (backend or customer server). Additionally, address verification is used to avoid accepting undeliverable mail - incoming mail is verified against either the backend or the customer server before accepting.
All information about domains/transports etc. is retrieved from a MySQL database, replicated from the backend to the MX frontends. That means, the frontends could actually use relay_recipient_maps to figure out for themselves, whether an address exists on the backend or not. The frontends would not have to bug the backend with verification probes. That is what I am aiming for.
However, the frontends do NOT know about valid recipients for the customers running their own mailservers. Hence, relay_recipient_maps cannot be used - as per postfix documentation, that would lead to all mail for those customers being rejected.
Is there a way to use relay_recipient_maps for backend domains while falling back to address verification for domains running their own mailserver?