From: Anthony Howe
Date: 2005-05-03 07:15:23 -0400
Subject: Re: milter-sender: excessive checks ?
More information..: http://www.milter.info/#Support
Jamie Jones wrote:
> My last message to the milter list came from a subdomain that
> accepts mail for any user at that domain.
> When I sent the mail, pop.snert.net rejected my message with
> a temporary 450 error, as expected.
> Then, the mail was passed to your next best MX, as expected.
> But shortly after (presumably when your secondary MX tried to send the mail),
> I received another callback, followed by another one 20 minutes later,
> presuambly when the secondary MX retried to send it....
> Not that it's a problem, but I thought these requests were cached ?
> I can see how the 10:54:17 entry appeared, as the IP address of the
> connecting machine would have been different, but as for the other...
The timeouts I use on pop.snert.net vary a lot, especially when I'm
testing milters. So for example CacheRejectTTL=900 and
CacheGcFrequency=20 keeps exercising the cache code tossing reject,
tempfail, and discard entries frequently, possibly faster than your
retry queue time. My secondary MXes are out side my control (though they
do use my software) and have more reasonable settings for production.
Once accepted though, they should be cached for about a week.
Also the cache is NOT shared between MX machines. To do that, you would
have to place the milter on one machine and use internet sockets from
two or more sendmail machines to the milter machine. This presupposes
that all the MXes are under your direct control, which is not my case.
Anthony C Howe +33 6 11 89 73 78
7116561 AIM: Sir Wumpus
in the mist of night / by the silent sea / a siren calls - Anthony
Copyright 2009, 2012 by SnertSoft. All rights reserved.