On 5/4/2005 19:32, "Paul" <paul@???> wrote:
>
> Hi Guys,
>
> Sorry if this has been answered in the past, I've tried a search in the
> archives. Exim 4.34 with a fairly standard configuration.
> We are relaying mail from one host to another and the destination host is
> running a greylist setup. No problem with that however our exim doesn't try
> for long enough and gives up too quickly.
> Below are the log entries I am seeing and also an extract from our exim
> retry configuration. I've tried something simple as double the values and it
> still didn't help and it just gave up after about 1 and 1/2 minutes.
>
> 2005-05-05 11:39:10 1DTVKY-0000i3-0u <= pop@??? U=pop
> P=local S=480 id=4279793E.mail23J1Q2XET@???
> 2005-05-05 11:39:11 1DTVKY-0000i3-0u SMTP error from remote mailer after
> RCPT TO:<Mrs.User@???>: host mx1.southern.net.au
> [100.100.100.2]: 450 <pop@???>: Sender address
> rejected: Service temporarily unavailable, please try later
> 2005-05-05 11:39:15 1DTVKY-0000i3-0u SMTP error from remote mailer after
> RCPT TO:<Mrs.User@???>: host mx1.southern.net.au
> [100.100.100.3]: 450 <pop@???>: Sender address
> rejected: Service temporarily unavailable, please try later
> 2005-05-05 11:39:18 1DTVKY-0000i3-0u SMTP error from remote mailer after
> RCPT TO:<Mrs.User@???>: host mx1.southern.net.au
> [100.100.100.4]: 450 <pop@???>: Sender address
> rejected: Service temporarily unavailable, please try later
> 2005-05-05 11:39:21 1DTVKY-0000i3-0u SMTP error from remote mailer after
> RCPT TO:<Mrs.User@???>: host mx1.southern.net.au
> [100.100.100.5]: 450 <pop@???>: Sender address
> rejected: Service temporarily unavailable, please try later
> 2005-05-05 11:39:23 1DTVKY-0000i3-0u == Mrs.User@???
> <Mrs.User@???> R=dnslookup T=remote_smtp defer (0): SMTP
> error from remote mailer after RCPT TO:<Mrs.User@???>:
> host mx3.southern.net.au [100.200.100.100]: 450
> <pop@???>: Sender address rejected: Service temporarily
> unavailable, please try later
>
>
> * * F,2h,15m; G,16h,1h,1.5; F,4d,6h
> *@TheDestination.com.au * F,8h,50m
>
> I tried other various combinations, also trying with exim -brt, none of
> which would match the seperate rule I put for the domain (or it's MX's as a
> test). Even doubling the default * rule values didn't help.
>
> Thanks
>
> Paul
>
The 15 minute retries for 2 hours should work with most greylisting
implementations (as Brian pointed out, the generic retry rule is the one
being used). The 8 hour ending of the specific rule is pretty quick
(although it's unlikely that ALL of the receiving MXs are down at once--it's
less unlikely than it would be were they on different /24s).
Trying all (or the "first" five) MX machines isn't going to help. Most
likely, they all work out of a common database, so you're probably recording
5 blocks in the one database.
You should look at a specific transport which instructs Exim to try only one
of the IPs for mx1.southern.net (and a router to get the messages there),
but that's not the issue. [It would save resources at both your end and
their end. But after you get one message through, the same sender/recipient
pair should sail through if it happens within 36 days (the usual setting,
since it allows things like first Monday of the month mail sendings to stay
in the database just long enough).]
None of that explains why Exim is giving up after the message's 13-second
lifetime in the queue, which is the actual problem you're facing
You should probably try tossing out the Exim hints databases. Or at least
run exinext against the domain to try to see see what Exim thinks is going
on.
--John
