[Exim] Clarification Sought (possibly a change for Exim4?)

Kezdőlap
Üzenet törlése
Válasz az üzenetre
Szerző: Paul Walsh
Dátum:  
Címzett: 'Exim Users'
Tárgy: [Exim] Clarification Sought (possibly a change for Exim4?)
I was just idly browsing through our reject log and noticed that on some
occasions sites that are bounced because they're in RBL also have an entry
through lack of reverse-DNS. This isn't a problem but I thought that if a
host were in RBL the mail would be dropped without further checks being
made, so I checked using exim -bh (aaa.bbb.ccc.ddd being an address found in
the reject log):

exim -d6 -bh aaa.bbb.ccc.ddd

*snip*

**** SMTP testing session as if from host aaa.bbb.ccc.ddd
**** Not for real!

host in host_lookup? no (option unset)
host in host_reject? no (end of list)
looking up host name for aaa.bbb.ccc.ddd
IP address lookup failed
host in host_reject_recipients? no (failed to find host name or lookup
deferred)
LOG: 0 MAIN
failed to find host name for [aaa.bbb.ccc.ddd]: permitted by +warn_unknown
host in rbl_hosts? yes (0.0.0.0/0)
checking RBL domain rbl-plus.mail-abuse.ja.net/reject
RBL lookup for aaa.bbb.ccc.ddd.rbl-plus.mail-abuse.ja.net succeeded
=> that means it is black listed at rbl-plus.mail-abuse.ja.net
LOG: 1 MAIN REJECT
recipients refused from [aaa.bbb.ccc.ddd] (RBL rbl-plus.mail-abuse.ja.net)
host in sender_unqualified_hosts? no (end of list)
host in receiver_unqualified_hosts? no (end of list)
host in helo_verify? no (option unset)
host in helo_accept_junk_hosts? no (option unset)


So it looks as though RBL checking is done AFTER host_reject and
host_reject_recipients. I wondered if there was a particular reason for
this as I thought that if the RBL check were done first (after all,
rbl_domains appears before host_reject in the config file) the mail could be
bounced and there would be no need for further DNS lookups.

Perhaps the order in which the tests are applied could be specified by
setting something like "rbl_reject_first"? The default would be to retain
the current order of processing unless rbl_reject_first were set

Paul