I'm receiving mail for a domain at a Postfix server on my VPS's static IP, then
forwarding it to a Postfix server on my office lan.
I've switched to connecting to the office lan's server @ an internal IP address
(192.168.2.13) over a VPN link that I've set up. That address is NOT reachable
directly from the 'net -- only over the VPN.
Mail sent through the VPS' IP is correctly forwarded and received at the
server. But once I stuck the VPN in the loop it's getting tagged as spam by
Amavis/Spamasssassin running on the office server.
The received message's headers have
X-Spam-Flag: YES
X-Spam-Score: 9.6
X-Spam-Level: *********
X-Spam-Status: Yes, score=9.6 tagged_above=-10 required=5
tests=[BAYES_00=-1.9, RDNS_NONE=1.5, SPF_HELO_FAIL=10]
autolearn=no autolearn_force=no
Received: from mx1-internal.MYDOMAIN.com ([127.0.0.1])
by localhost (mx1-internal.MYDOMAIN.com [127.0.0.1])
(amavisd-new, port 10024)
with ESMTP id CoqYh_JBKbcT for <[email protected]>;
Sat, 16 May 2015 18:13:17 -0700 (PDT)
Received: from mx1.MYDOMAIN.com (unknown [10.0.0.1])
(using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits))
(No client certificate requested)
by mx1-internal.MYDOMAIN.com (Postfix) with ESMTPS id
BABB5101D7B
for <[email protected]>; Sat, 16 May 2015 18:13:16 -0700
(PDT)
Received: from c.mail.sonic.net (c.mail.sonic.net [64.142.111.80])
by mx1.MYDOMAIN.com (Postfix) with ESMTP id D195061EC3
for <[email protected]>; Sat, 16 May 2015 18:13:15 -0700
(PDT)
The main culprit looks like it's the
SPF_HELO_FAIL=10
Probably because the receiving Postfix server checks the HELO @ IP = 10.0.0.1,
if I understand how the HELO checks work
Received: from mx1.MYDOMAIN.com (unknown [10.0.0.1])
Where 10.0.0.1 is the VPN endpoint on the VPS side.
The mail logs on the Office side show
May 16 18:42:55 mx1-internal amavis[6073]: (06073-02) Passed SPAMMY
{RelayedTaggedInbound}, [10.0.0.1]:53597 [208.201.242.19] <[email protected]> ->
<[email protected]>, Queue-ID: BABB5101D7B, Message-ID:
<[email protected]>, mail_id: CoqYh_JBKbcT, Hits: 9.6,
size: 1536, queued_as: 0E73B101D89, 7202 ms
I think I need to fix or ignore that check - since the relay step from my VPS
to my Office server isn't relevant to checking.
What I DON'T KNOW, though, is where I fix this. In Postfix config, either on
the VPS or Office side, or in Amavis/Spamassassin. I'm not convinced that
because amavis is reporting on it that it's an amavis issue. Might be ...
Where do I "ok" this hop so that it's not scored, or even considered in the
scoring?
