Long explanation. Sorry.
I have a mailman mailing list machine running Exim called
pirate.real-time.com, which I recently turned on VERP.
I have a backup MX host called mx1.real-time.com running Exim which does
sender and recipient verification.
require verify = sender/callout=90s/check_postmaster
accept domains = +relay_domains
endpass
message = MX-relay attempt, recipient verification failed
verify = recipient/defer_ok/callout=90s,defer_ok
I have a "smtp proxy" running Exim called gatekeeper.tanners.org.
I took pirate down for maintenance. All mail spooled up on mx1 just fine. When
pirate came back online, mx1 started to feed the queued mail back to pirate.
This is where the problem started.
Pirate's logs show:
2003-11-03 00:33:47 H=localhost.localdomain (pirate.real-time.com) [127.0.0.1]
sender verify fail for <tanner-list-bounces+bob=tanners.org@???>:
response to "RCPT TO:<tanner-list-bounces+bob=tanners.org@???>" from
gatekeeper.tanners.org [65.165.40.73] was: 550 unknown user
2003-11-03 00:33:47 H=localhost.localdomain (pirate.real-time.com) [127.0.0.1]
F=<tanner-list-bounces+bob=tanners.org@???> rejected RCPT
<bob@???>: Sender verify failed
I think this started when I added queue_only = true to pirate's exim.conf
On gatekeeper.tanners.org, I see in the log this:
2003-11-03 00:32:57 H=pirate.real-time.com [208.20.202.13] F=<> rejected RCPT
<tanner-list-bounces+bob=tanners.org@???>: Unrouteable address
bob@??? is a valid address on gatekeeper.
Why is the sender verify failing?
It's like gatekeeper doesn't understand VERP email addresses.
Is my recipient verification rule incorrect?
Any help would be appreciated.
--
Bob Tanner <tanner@???> | Phone : (952)943-8700
http://www.mn-linux.org, Minnesota, Linux | Fax : (952)943-8500
Key fingerprint = AB15 0BDF BCDE 4369 5B42 1973 7CF1 A709 2CC1 B288