Hej,
once again I'm experiencing an annoying problem related to greylisting
over here, and I think this may be considered a bug within exim. (If I
figured out everything correctly.)
The MX records for rtr.at point to mail.rtr.at(5) and
mailbackup.inode.at(10), the former using greylisting, the latter - of
course - not. Now, when exim is trying to deliver a message to
user@??? it gets a temporary error from mail.rtr.at and then delivers
the message to mailbackup.inode.at. An entry is written to the retry DB,
stating user@??? failed on mail.rtr.at.
Then, if some days later another message gets routed to user@???, it
will again fail temporarily on mail.rtr.at. But now, if the difference
between the time of the first and the current failure is greater than
the maximum retry time, exim will generate a permanent error, without
trying to deliver to mailbackup.inode.at. The retry data then looks
something like this:
R:user@??? -44 13133 SMTP error from remote mailer \
after RCPT TO:<user@???>: host mail.rtr.at [81.223.31.36]: \
451 4.7.1 Greylisting in action, please come bac
28-Apr-2006 19:26:45 03-May-2006 10:53:19 03-May-2006 16:53:19 *
Retry config is quite simple:
* * G,4h,10m,1.5; F,16h,2h; F,4d,6h
Besides I don't like the fact that retry time is calculated in
refrerence to the first failure in that case, (this can be tweaked a
little by lowering retry_data_expire) IMHO exim should never bounce a
mail after trying only one of two MX hosts.
What do you think?
The corresponding host is running exim 4.50 (Debian/sarge) but I
couldn't find anything related to this in the Changelogs.
lg,
daniel
