2

I'm using transport table to specif the nexthop for specific domain:

Host A:

/etc/postfix/transport
domain.com     smtp:[mail.host2.com]:25

Host A verifies local recipients with lookup table:

local_recipient_maps = proxy:unix:passwd.byname $alias_maps
unknown_local_recipient_reject_code = 550
  • but I wonder if there is a way to verify remote recipients on Host B from Host A at SMTP time to avoid NDR to forged emails.

Update: Mon 28 Jul 13:49:01 BST 2014:

It looks like Exim can do recipient email verification via LMTP so I wonder if Postfix can do the same?

# Perform recipient verification
deny message = Recipient verification failed, Non-existent mailbox
log_message = Recipient verification failed
!verify      = recipient/callout=2m,use_postmaster,defer_ok

1 Answer 1

2

There is no such generally-valid way. Not only would such a technique be of immense use to spammers, but in addition you can't know what happens to mail once it passes through a company's externally-advertised mail gateway.

Perhaps it stays in spool on that box while it's checked for spam and viruses, then is delivered to an internal server for final delivery and reading by the recipient; that's a common enough setup. The relay box doesn't have the faintest idea which users are really valid - it just scrubs content and passes it on - so even if there was a way to ask it, it couldn't tell you.

Edit in light of update: that's not going to help you. The article you link to shows a relay box that accepts mail for inbound final delivery, inside an organisation, and makes it aware of user existence/nonexistence on the point of final delivery via LMTP callout.

In my example above, LMTP delivery could be used on the (hypothetical) remote relay system to enable it to know about which recipients would end up being a valid user at final inbound delivery. But you don't have any way of forcing them to enable that on their relay box, and unless you do, their relay box doesn't know what to tell you when you try to validate recipients.

Basically, there are lots of techniques to validate recipients inside the trust boundary of an organisation, but (fortunately) there are no generally-valid ones that work across the internet-at-large.

2
  • Please see my update.
    – HTF
    Jul 28, 2014 at 12:49
  • Please see mine.
    – MadHatter
    Jul 28, 2014 at 14:21

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .