You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Sergei Gerasenko <ge...@publicschoolworks.com> on 2006/05/20 21:46:33 UTC
spamc/spamd/bayes
Hello,
I started piping my mail through SA a couple of months ago and I've
been diligently marking messages as spam for the bayes subsystem. Then I
noticed that neither the headers of messages nor the analysis reports
have anything about bayes rules.
I'm running exim and here's what I have in the config file for the
spamcheck transport:
spamcheck:
debug_print = "T: spamassassin_pipe for $local_part@$domain"
driver = pipe
command = /usr/sbin/exim4 -oMr spam-scanned -bS
use_bsmtp
transport_filter = /usr/bin/spamc
home_directory = "/tmp"
current_directory = "/tmp"
user = Debian-exim
group = Debian-exim
return_fail_output
message_prefix =
message_suffix =
The problem is that exim runs under Debian-exim (I'm running Debian)
and the bayes db is under /root/.spamassassin.
I tried to add "-u root" to /usr/bin/spamc just to test it, and while
the bayes rules were consulted (or so the headers of a message said), I
got "autolearn=failed". I think it's because SA was trying to update the
bayes database under /root/.spamassassin.
So, my question is, what's the correct way of doing this so that the
Bayes db is system wide?
I'm sorry if this has been asked before. If so, would somebody please
point me in the right direction?
Thanks for any suggestions,
Sergei
Re: spamc/spamd/bayes
Posted by Michael Monnerie <mi...@it-management.at>.
On Montag, 22. Mai 2006 17:46 Sergei Gerasenko wrote:
> Since 90% of users use outlook/outlook express, how do they report
> their spam to you? Put differently, how do you extract spam out of
> outlook and outlook express mailboxes?
IMAP Account. Or see http://www.olspamcop.org/
mfg zmi
--
// Michael Monnerie, Ing.BSc ----- http://it-management.at
// Tel: 0660/4156531 .network.your.ideas.
// PGP Key: "lynx -source http://zmi.at/zmi3.asc | gpg --import"
// Fingerprint: 44A3 C1EC B71E C71A B4C2 9AA6 C818 847C 55CB A4EE
// Keyserver: www.keyserver.net Key-ID: 0x55CBA4EE
Re: spamc/spamd/bayes
Posted by Dan <a...@patnode.net>.
> Since 90% of users use outlook/outlook express, how do they report
> their spam to you? Put differently, how do you extract spam out of
> outlook and outlook express mailboxes?
Sergei,
This may help:
http://daesoft.com/SpamSource/
Re: spamc/spamd/bayes
Posted by Sergei Gerasenko <ge...@publicschoolworks.com>.
> We use site-wise bayes here too. While users can report FN's and FP's,
> IT staff reviews the submissions prior to actual learning. This prevents
Since 90% of users use outlook/outlook express, how do they report
their spam to you? Put differently, how do you extract spam out of
outlook and outlook express mailboxes?
Re: spamc/spamd/bayes
Posted by Michael Monnerie <mi...@it-management.at>.
On Montag, 22. Mai 2006 17:28 Bret Miller wrote:
> Either that or the other 95% of users get no spam ever.
*bruahaha* I just spit that nice coffee over my keyboard...
mfg zmi
--
// Michael Monnerie, Ing.BSc ----- http://it-management.at
// Tel: 0660/4156531 .network.your.ideas.
// PGP Key: "lynx -source http://zmi.at/zmi3.asc | gpg --import"
// Fingerprint: 44A3 C1EC B71E C71A B4C2 9AA6 C818 847C 55CB A4EE
// Keyserver: www.keyserver.net Key-ID: 0x55CBA4EE
RE: spamc/spamd/bayes
Posted by Bret Miller <br...@wcg.org>.
> On Montag, 22. Mai 2006 01:12 Sergei Gerasenko wrote:
> > But I'm reading everywhere that it's not a paricularly good idea.
>
> There's not a single answer to whether which method is best. I use a
> sitewide bayes, and it works good.
>
> sitewide:
> + spam learned helps all users
> + good when trained with 100% correct spam/ham
> - dangerous when learning false spam/ham
>
> user:
> + each user can train themselves
> - users most often don't train good, or not at all, or false (YMMV)
> - performance
> - disk space
We use site-wise bayes here too. While users can report FN's and FP's,
IT staff reviews the submissions prior to actual learning. This prevents
people from learning various e-mail lists they've signed up for as
SPAM-- we just send the report back and say, try unsubscribing first.
The approach has worked fairly well for us here. The number of users
that actually report anything is probably around 5%, so I'd say that a
per-user system would be less effective for our users. (Either that or
the other 95% of users get no spam ever.)
Bret
Re: spamc/spamd/bayes
Posted by Michael Monnerie <mi...@it-management.at>.
On Montag, 22. Mai 2006 01:12 Sergei Gerasenko wrote:
> But I'm reading everywhere that it's not a paricularly good idea.
There's not a single answer to whether which method is best. I use a
sitewide bayes, and it works good.
sitewide:
+ spam learned helps all users
+ good when trained with 100% correct spam/ham
- dangerous when learning false spam/ham
user:
+ each user can train themselves
- users most often don't train good, or not at all, or false (YMMV)
- performance
- disk space
mfg zmi
--
// Michael Monnerie, Ing.BSc ----- http://it-management.at
// Tel: 0660/4156531 .network.your.ideas.
// PGP Key: "lynx -source http://zmi.at/zmi3.asc | gpg --import"
// Fingerprint: 44A3 C1EC B71E C71A B4C2 9AA6 C818 847C 55CB A4EE
// Keyserver: www.keyserver.net Key-ID: 0x55CBA4EE
Re: spamc/spamd/bayes
Posted by Sergei Gerasenko <ge...@publicschoolworks.com>.
All right, I figured out the problem below by just adding bayes_path to
/etc/spamassassin/local.cf. So, now I have a shared bayes db. But I'm
reading everywhere that it's not a paricularly good idea. So, the
question is, how do I with the same setup (exim, etc), make spamd setuid
to the recipient of each message and read that user's bayes db?
Judging by the debug output of spamd, it just switches to the user that
called spamc and that's the end of the story.
Is replacing "user = Debian-exim" with "user=$localpart" a good idea? I
guess this is a question for exim users.
Just curious if anybody has a quick tip on this. All right, off to read
the exim manual...
On Sat, May 20, 2006 at 03:46:33PM -0400, Sergei Gerasenko wrote:
> Hello,
>
> I started piping my mail through SA a couple of months ago and I've
> been diligently marking messages as spam for the bayes subsystem. Then I
> noticed that neither the headers of messages nor the analysis reports
> have anything about bayes rules.
>
>
> I'm running exim and here's what I have in the config file for the
> spamcheck transport:
>
> spamcheck:
> debug_print = "T: spamassassin_pipe for $local_part@$domain"
> driver = pipe
> command = /usr/sbin/exim4 -oMr spam-scanned -bS
> use_bsmtp
> transport_filter = /usr/bin/spamc
> home_directory = "/tmp"
> current_directory = "/tmp"
> user = Debian-exim
> group = Debian-exim
> return_fail_output
> message_prefix =
> message_suffix =
>
> The problem is that exim runs under Debian-exim (I'm running Debian)
> and the bayes db is under /root/.spamassassin.
>
> I tried to add "-u root" to /usr/bin/spamc just to test it, and while
> the bayes rules were consulted (or so the headers of a message said), I
> got "autolearn=failed". I think it's because SA was trying to update the
> bayes database under /root/.spamassassin.
>
> So, my question is, what's the correct way of doing this so that the
> Bayes db is system wide?
>
> I'm sorry if this has been asked before. If so, would somebody please
> point me in the right direction?
>
> Thanks for any suggestions,
> Sergei