Hi Luca,
I don't have an answer, but I'm also interested in this and am adding
more info around my experience in case this is helpful. I have
experienced it in two distinct patterns:
1. (ignorable) I get it at midnight when logrotate happens. Meh, I don't
care about this one!
2. On one of my servers I got this all the time for a certain type of
delivery. I spent ages trying to unpick why and did not get anywhere.
Like you, the entries were correctly written to the log file that it
apparently could not write to... BUT, the mail was delivered OK, but the
exim crash/error caused CiviCRM (generating the mail) to believe the
mail had *not* been sent OK.
I ran pretty much identical exim config on that server as others, and
all my servers run debian, too - but only one was affected. Last time I
literally started from scratch on a different VPS and did not encounter
the problem again. I had posted to this list (16 Mar 2022) and received
one reply which offered that it could be a timing/log rotation error,
which it was not - I could easily reproduce the fault at any time.
Other info in case it is relevant: Since this became rare for me again,
I've noticed that it was often, but not always, to microsoft domains:
hotmail, outlook, live, and often it came after a
> (send) The TLS connection was non-properly terminated.
...error on the previous log line at the same timestamp.
Rich Lott (they/them <https://en.pronouns.page/they/them/themself>)
--
## subscription configuration (requires account):
## https://lists.exim.org/mailman3/postorius/lists/exim-users.lists.exim.org/
## unsubscribe (doesn't require an account):
## [email protected]
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/
