Steffen Nurpmeso via Postfix-users: > |I might add, in case that it isn't obvious from the above, that there > |is no deduplication across multiple queue files. > | > |If the mailing list generates multiple submissions into Postfix, > |then duplicates will be suppressed only within the same queue file. > | > |If you really want useful information from me, then I need to know > |what route the duplcated recipients took in terms of queue files. > > Well, please excuse me, i do not understand. No, it is not > obvious. I mean, i have shown the sendmail(1) invocation with two > recipients -- granted this is all local --, and the one is > expanded, so we effectively have the same address twice, and that > is then "delivered" via local, it is actually an alias that runs > a program, and that is run twice.
That is insufficient data to reproduce the observation. > Whereas i would expect something like > > </dev/null s-nail -:/ -Sstealthmua -Smta=test \ > -X 'alias [email protected] [email protected]' \ > -s hi [email protected] [email protected] > From steffen Wed May 6 22:45:29 2026 > Date: Wed, 06 May 2026 22:45:29 +0200 > To: [email protected] > Subject: hi What are the envelope recipients in the queued message? Maybe you can add a temporary header_checks rule with a HOLD action. Then you can inspect the envelope recipients with the mailq command (or postqueue -p). Can you show those recipients? Which recipients are the result from virtual_alias_maps mapping? Are any of these envelope recipients subject to aliasing with Postfix alias_maps? Once the message is moved out of the 'hold' queue with 'postsuper -H', the queue manager will log the number of recipients as: qmgr[pid] ...., nrcpt=number, ... (maybe 'postqueue -f' is needed to speed this up) The final recipients are then logged by the local delivery agent. Can you show those records? Of particular interest are the PIDs of the local daemon process and course the final recipient addresses. Wietse _______________________________________________ Postfix-users mailing list -- [email protected] To unsubscribe send an email to [email protected]
