On Mon, 27 Oct 2014 04:09:09 +0100
"li...@rhsoft.net" <li...@rhsoft.net> wrote:

> 
> Am 27.10.2014 um 03:38 schrieb ferriswheel:
> > On Mon, 27 Oct 2014 02:37:00 +0100
> > "li...@rhsoft.net" <li...@rhsoft.net> wrote:
> >>
> >> Am 27.10.2014 um 02:22 schrieb ferriswheel:
> >>> currently using postifix version 2.5.1
> >>> how do i configure lmtp to keep a  message queue when the outgoing link 
> >>> is down
> >>
> >> uhm there is nothing to configure
> >>
> >> a message keeps queued until it is delivered and even if the lmtpd
> >> daemon crashs until it responded with success it still get queued and
> >> maybe delivered more than once (what is even logged in that case)
> >>
> >> 2.5.1 is stoneold but that behavior i guess is default for decades since
> >> lmtp is practically the same as smtp
> >
> > the postfix lmtp process in this situation has nothing to connect to, as 
> > the destination is not listening.
> 
> so what - that happens all day long in case of smtp
> 
> > as shown in this sanitised log entry the qmgr process is told to give up as 
> > status=expired
> >
> >     Oct 20 07:45:52 pilot mail_pilot/lmtp[5624]: connect to X.X.X.X:22424: 
> > No route to host
> >     Oct 20 07:45:52 pilot mail_pilot/lmtp[5624]: 579032410732: 
> > to=<m...@example.com>, relay=none, delay=3.1, delays=0.04/0/3/0, dsn=4.4.1, 
> > status=deferred (connect to Y.Y.Y.Y:22424: No route to host)
> >     Oct 20 07:45:52 pilot mail_pilot/qmgr[4951]: 579032410732: from=<>, 
> > status=expired, returned to sender
> >     Oct 20 07:45:52 pilot mail_pilot/qmgr[4951]: 579032410732: removed
> 
> status=expired?
> 
> sounds like you have f**d up "maximal_queue_lifetime" which is normally 
> 5 days and that is why you should always post "postconf -n" output as 
> you where directed to debug-readme in the first reply
> 
> [root@rh:~]$ postconf -n maximal_queue_lifetime
> maximal_queue_lifetime = 3d
> 
> [root@rh:~]$ postconf -d maximal_queue_lifetime
> maximal_queue_lifetime = 5d


hello,


thanks for your reply.


yes, that was the problem.

maximal_queue_lifetime and bounce_queue_lifetime were set to '0'


regards

john

Reply via email to