On Thu, 18 Sep 2003, Dennis Skinner wrote:
> server1.domain.com * F,3h,5m; G,16h,1h,1.5; F,4d,6h
> * * F,3h,15m; G,16h,1h,1.5; F,4d,6h
>
> >From the exim docs on retry rules, it sounds like these settings should
> cause server1's retry to be 5mins and 15 for everything else for the
> first 3 hours.
Yes, but you have to understand how the retry rules work.
> Will this retry rule do what I want (ie if problem sending to server1,
> only wait 5mins before trying again for messages to bluefrog.com)?
Maybe. :-) It depends on how often you start queue runners. What the
retry rules do is to specify how Exim computes the "earliest next try
time" after a temporary error. These times are not guarantees of
delivery attempts. They are hints to the delivery processes started by
queue runner processes.
Such a delivery process looks at a message and asks the question "is it
time to try this delivery again?" for each undelivered address.
Therefore, the actual time of next try depends both on the retry rules
and on the frequency of starting queue runners.
If you start a queue runner every 15 minutes (a typical configuration)
then, on average, a queue runner will look at a failing message 7.5
minutes after it has failed. So a retry time of 5 minutes as opposed to
15 minutes will have some effect, but it doesn't mean that Exim will
"only wait 5mins before trying again".
Philip
--
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
