1

I have a problem with e-mails sent to a particular domain (alice.it). Usually the mail is corrected sent and in my postfix log I have this:

06104180AF: to=<**********@alice.it>, relay=mx.tim.it[34.141.161.132]:25, delay=200, delays=0/200/0.14/0.12, dsn=2.0.0, status=sent (250 <62CC171918270D1E> Mail accepted)

But sometimes, the mails are not sent and in the logs I receive this error:

11F7817D05: to=<**********@alice.it>, relay=none, delay=59, delays=0/59/0/0, dsn=4.0.0, status=deferred (delivery temporarily suspended: host mx.tim.it[34.141.161.132] refused to talk to me: 421 mta-in-01.tim.it Service not available - too busy)

The difference is in the field "relay" that in the first case is "mx.tim.it". In the second is "none".

Also in my postfix configuration, inside the file /etc/main.cf i have:

transport_maps = hash:/etc/postfix/rl_transport

And in /etc/postfix/rl_transport, I have:

alice.it slowest:

In the /etc/postfix/master.cf I have:

slowest unix - - n - - smtp -o syslog_name=postfix-slowest

So, all the mail to alice.it domaini sould be tagged with "postfix-slowest", but the ones that have "relay=none" are tagged with "postfix". The others (the ones with relay=mx.tim.it") instead are correctly tagged with "postfix-slowest".

How is it possible?

There is something that I can do, or is a problem of the receiver (alice.it)?

Thanks a lot

Stefano

Stefano
  • 11
  • 1
  • The delays field suggests that *both* of these messages were not sent straight away. Search for the message ID (e.g. `06104180AF` / `11F7817D05`) or relay, and I expect 200 / 59 seconds earlier you will find another attempt for those messages. – anx Sep 22 '22 at 16:48
  • You can move those messages to a separate transport as you have done, but that mostly helps you pay less attention to the unavoidable delays. They have been unable to allocate sufficient resources (or to prioritize legitimate connections) for years and delays are simply to be expected for that operator. – anx Sep 22 '22 at 17:28
  • Hi anx and thx for the reply. I have searched for the 2 ID. For the first one I have not found other occurrences. For the second, I have found a second occurrence after 8 minutes where the message has been sent correctly: ```11F7817D05: to=<*************@alice.it>, relay=mx.tim.it[34.141.161.132]:25, delay=544, delays=505/39/0.14/0.09, dsn=2.0.0, status=sent (250 <629DCC40247D2E7F> Mail accepted)``` The second message has been tagged correctly with postfix-slowest. It's very strange. Why the first message is not tagged with slowest, and the second attempt yes? – Stefano Sep 23 '22 at 07:20
  • I can neither find it documented nor understand the code sufficiently, but my working theory is: the missing transport-specific log prefix is caused by some detail in *how postfix caches per-site unavailability* (why did postfix not immediately deliver your first sample message? possibly because it already knew that the destination had recently greeted with 421. not necessarily while attempting to deliver *that specific* message, though). – anx Sep 23 '22 at 18:46

0 Answers0