Re: [exim] "retry time not reached for any host after a long…

Páxina inicial
Borrar esta mensaxe
Responder a esta mensaxe
Autor: Christiaan den Besten
Data:  
Para: exim-users
Asunto: Re: [exim] "retry time not reached for any host after a longfailureperiod" (2)
Yes, just changed it to:

---
begin routers

outgoing:
  driver                = accept
  domains               = ! +local_domains
  transport             = remote_smtp
  no_more


incoming:
  driver                = accept
  transport             = local_queue
  check_local_user


begin transports

local_queue:
  driver                = appendfile
  file                  = /dev/null


remote_smtp:
  driver                = smtp
---


Now all recipient checks should be able to succeed without errors :)

Daemons run with the following parameters, I will kill the -q5m for the
incoming queue .....

---
Debian-  12592  0.0  0.2  8928 2516 ?        Ss   16:43   0:00 
/usr/sbin/exim4 -C /etc/exim/exim4-in.conf -oP 
/var/run/exim4/exim-in.pid -bd -q5m
Debian-  12595  0.0  0.2  8880 2448 ?        Ss   16:43   0:00 
/usr/sbin/exim4 -C /etc/exim/exim4-out.conf -oP 
/var/run/exim4/exim-out.pid -q5m
---


bye,
Chris

----- Original Message -----
From: "Philip Hazel" <ph10@???>
To: "Christiaan den Besten" <chris@???>
Cc: <exim-users@???>
Sent: Wednesday, October 13, 2004 4:54 PM
Subject: Re: [exim] "retry time not reached for any host after a
longfailureperiod" (2)


> On Wed, 13 Oct 2004, Christiaan den Besten wrote:
>
>> I use Mailscanner for mail filtering, so the 'input' exim daemon has
>> 'queue
>> only' set in the configuration. But exim requires to have a 'valid'
>> router and
>> transport anyway .... ( even when you have queue_only enabled ).
>
> I presume that is because you are doing recipient verification?
>
>> To make exim
>> happy I made the following router:
>>
>> smarthost:
>> driver = manualroute
>> domains = *
>> transport = remote_smtp
>> route_list = * 1.2.3.4
>> no_more
>>
>> ( localhost does not work, for exim than thinks it should forward the
>> message
>> to the same host and suspects a loop )
>
> Two comments: (1) Since that always accepts, why bother with the
> recipient verification? Just turn it off in the ACL. (2) If you just
> want a router/transport that accepts, try this:
>
> xxxx:
> driver = accept
> transport = yyy
>
> yyy:
> driver = appendfile
> file = /dev/null
>
> But, reading further, I'm not sure this is right, so don't do it till
> you understand the rest...
>
>> It should simply accepts all mail (yes, should have used an 'accept'
>> router,
>> juist build that one :) ) and should never realy be used because the
>> daemon
>> should just (and does) queue the message.
>>
>> But, exim -does- check if it can reach the (nonexistent) host. So its
>> gives an
>> error once every 5 days ;) ....
>
> That is actually strange. It should not be trying to do any deliveries
> when queue_only is set. Have you remembered to start the 'input' daemon
> with just -bd? If you have used '-q10m' or similar, it's going to be
> trying to do deliveries, and you'll get this effect....
>
>> Is there any particular reason for exim to check for 'valid'
>> transports/routers in a queue_only mode?
>
> It will verify if you have used 'verify = recipient' in an ACL (or
> 'verify = sender') but it should not try to deliver.
>
>
> -- 
> Philip Hazel            University of Cambridge Computing Service,
> ph10@???      Cambridge, England. Phone: +44 1223 334714.
> Get the Exim 4 book:    http://www.uit.co.uk/exim-book

>
>