Exim+MailScanner - not always queuing mail on "in" side

[Tim Bishop]

On Mon, Jun 02, 2003 at 02:01:07PM +0100, Tony Finch wrote:
> Tim Bishop <tim-lists at BISHNET.NET> wrote:
>
> >The problem is that occasionally (but not always) locally generated
> >messages such as cron output don't get deferred by the incoming exim.
> >From my understanding of Exim it seems that it's ignored the queue_only,
> >then tried to defer it. Then, for some reason, it's decided to fail
> >the message.
>
> This is probably because the hints database for the incoming exim says
> that addresses have been failing for such a long time that they bounce
> immediately. You need to check that the spool directory configurations
> for the incoming and outgoing exims are correct and that they are being
> run with the correct commands, and check that /var/spool/exim_incoming/db
> is empty (as it should be if the queue_only option is working).

I did have a retry file in the exim.in/db directory - which was
causing the bouncing. However, I think I know how this got there.

Cron on FreeBSD runs sendmail (well, exim) with the -odi flag, which
causes a second exim process to attempt delivery - even with the
queue_only option switched on. Turning off this flag seems to have
gone part way to fixing this... but it's still not entirely happy.

It'd be nice if exim had a queue_only_always flag which would make it
queue every time, and not permit anything else.

Tim.

--
Tim Bishop
http://www.bishnet.net/tim
PGP Key: 0x5AE7D984