On Thu, Dec 11, 2025 at 01:55:16AM +0100, Gerald Galster via Postfix-users
wrote:
>
> > Viktor Dukhovni via Postfix-users <[email protected]>:
> >
> > On Thu, Dec 11, 2025 at 12:55:53AM +0100, Gerald Galster via Postfix-users
> > wrote:
> >
> >> I'm relating to the modified log message in general:
> >
> > You failed to read Wietse's original carefully. The only proposed
> > change is to log **failure to create** a bounce message queue file with
> > the original queue id as the correct context. This does not change the
> > logging of normal (successful) bounce processing.
>
> Thanks for clarifying that.
>
> So for successful deliveries nothing changes:
>
> > postfix/bounce[]: 4dRQzN6pXBzVHYR: sender non-delivery notification:
> > new-queue-id
>
> And the error case would look like this:
>
> > postfix/bounce[]: 4dRQzN6pXBzVHYR: sender non-delivery notification:
> > message content rejected
>
> How shall a parser best distinguish a queue id from a possibly variable error
> message?
The message could be subtly different, perhaps one of:
postfix/bounce[]: <queue-id>: sender non-delivery notification message
content rejected
postfix/bounce[]: <queue-id>: sender non-delivery message content rejected
postfix/bounce[]: <queue-id>: rejected: sender non-delivery notification
message
...
Of course the real solution is to NOT configure your MTA to reject
bounces. And perhaps send header-only boucnes, to reduce the chance
that the body will trigger filters.
If your system reports this sort of message, the problem isn't its
syntax, it is the fact this is happening at all.
--
Viktor. 🇺🇦 Слава Україні!
_______________________________________________
Postfix-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
