1 |
I say OT because it's my understanding of DKIM that lets me down here, not Gentoo. I'm |
2 |
just not sure who to ask or even if it could be something Gentoo related. |
3 |
|
4 |
I've recently updated my postfix home mail server to use amavis-new for virus and spam |
5 |
filtering rather than procmail/spamassassin. |
6 |
|
7 |
It seems to be working well and I've also enabled some other goodies like DKIM signing |
8 |
and verification. I haven't confirmed signing is working yet, so maybe a side effect |
9 |
of this email is that someone can confirm this for me ;) |
10 |
|
11 |
The main query I have is that a lot of the mail I get, in this case from various |
12 |
mailing lists, appears to failed DKIM verification. |
13 |
|
14 |
For example, several of the posters on this list are DKIM signing their mail either as |
15 |
part of gmail policy (or another big provider) or personal intent. Something in the |
16 |
region of 50% of signed mail on this list contains headers such as: |
17 |
|
18 |
Authentication-Results: genesis.genestate.com (amavisd-new); dkim=softfail |
19 |
(fail, message has been altered) header.i=@gmail.com |
20 |
Authentication-Results: genesis.genestate.com (amavisd-new); domainkeys=softfail |
21 |
(fail, message has been altered) header.from=xxxxxx@×××××.com |
22 |
|
23 |
Whereas the rest looks like this: |
24 |
|
25 |
Authentication-Results: genesis.genestate.com (amavisd-new); dkim=pass |
26 |
header.i=@gmail.com |
27 |
Authentication-Results: genesis.genestate.com (amavisd-new); domainkeys=pass |
28 |
header.from=xxxxxx@×××××.com |
29 |
|
30 |
Now I find it unreasonable to assume that 50% of the mail I receive is being actively |
31 |
tampered with, so it must be something getting twisted out of shape. All I'm trying to |
32 |
discover is whether it's something at my end that I need to fiddle with. I followed a |
33 |
few different guides to piece my setup together so it's quite possible I've overlooked |
34 |
or misconfigured something. |
35 |
|
36 |
If anyone knows about DKIM and might be able to shed a light on this, I'd love to |
37 |
hear. It's not a big problem, just a puzzle I'm interested in. |
38 |
|
39 |
Thanks |
40 |
|
41 |
Matt Harrison |