Hi,
We have a redirect router in Exim 4.01 that does a ${lookup{$local_part}
from an /etc/aliases style alias file for forwarding. The only difference
is that this particular alias file is under control of one of our customers.
Anyway, one of the entries had a syntax error in it that was causing
an error of:
== lee_lewis@??? R=virtual_aliasfile defer (-17): error in redirect data: malformed address: or l may not follow efg@??? in "efg@??? or l"
LOG: retry_defer MAIN
My concern is that our retry configuration consists of a single line of:
* * F,2h,15m; G,16h,1h,1.5; F,4d,4h
Yet, this message was retrying over and over for the past 7 days. From
the Exim spec, it appears that it should have hit an ultimate-message-timeout
and bounced back to sender. If it had done so, the user would have probably
noticed and ultimately fixed the problem.
My question, I guess, is whether or not routing failures adhere to
ultimate message retry timeouts? If so, would a syntax routing error
like this also be candidate?
Also, when I would manually do a "exim -q", it would *always* attempt
rerouting of the message, rather than saying "retry timeout not reached"
like it does for most messages. Further, an exim_dumpdb on the retry
database would never show ANY entry for the local-part/domain in question.
Not sure if this is my problem or an Exim problem, but I would have
thought that no message would have ever been allowed to be on the
queue more than 4 hours based upon our retry rule config above.
Regards,
Dean Brooks
dean@???
