On Fri, May 22, 2026 at 02:57:07PM +0530, Rajesh Mishra via Postfix-users wrote:
> TCP lookup tables (tcp:) send the full recipient address to the daemon exactly
> once. They do not attempt the multi-form lookups (user+ext@domain,
> user@domain,
> @domain) that indexed maps perform. This means my TCP daemon on port 9926 is
> solely responsible for handling plus-addressing, catchalls, and alias-domain
> canonicalisation — Postfix will not retry with a stripped or rewritten key.
Nothing precludes A TCP daemon, a PCRE table, or a policy service from
examining addresses parts and making a decision based on its various
components.
> One thing I want to confirm: setting virtual_mailbox_maps = <empty> while
> keeping smtpd_reject_unlisted_recipient = yes does NOT give us what we want —
> it disables the recipient check entirely rather than rejecting unknowns. So
> an
> empty map is not the right answer.
There's no reason to solve non-problems. Let a "reduntant" lookup
happen, it is not redundant for others, and harmless in your case.
> --- Questions ---
>
> 1. For Design A: is there any scenario where Postfix would still consult
> virtual_mailbox_maps for something other than
> smtpd_reject_unlisted_recipient
> validation? For example during cleanup, queue re-injection, or virtual(8)
> delivery triggered by some other path?
The virtual(8) delivery agent uses the table to determine where the
mailbox is stored.
--
Viktor. 🇺🇦 Слава Україні!
_______________________________________________
Postfix-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
