You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by "Steve [Spamassasin]" <sp...@shic.co.uk> on 2005/09/12 11:27:17 UTC
Very simple user query...
I'm using spamassassin (Razor, Pyzor, DCC) and procmail to filter all my
mail on my (Gentoo) linux-server, to which I connect from a number of
Windows (XP/2000) machines using Mozilla Thunderbird to access my
(dovecot) IMAP folders on the linux server. I configured spamassassin
to use "Rulesdujour" and to regularly update those rules - and I was
very happy... at least 99.99% of spam was correctly marked with only one
incident of false positives (for which spamassasin wasn't entirely to
blame.) in several months.
Lately I've been less lucky - only ~99% of my spam is marked as such...
which sounds good but the remaining 1% gives me up-to a dozen bogus
messages each day... which is frustrating. To the naked eye the missed
spam is obviously spam - but typically the only significant rule it
triggers is the Bayesian rule... As I've stuck to the default settings
this alone is insufficient to identify a mail as spam.
I'm left with several questions...
* Is there somewhere where I can report spams which aren't caught by
the default configuration in order to feed-back into future
improvements?
* Is there an easy way to report spam explicitly to the checksum
services (Razor/Pyzor/DCC)?
Any other suggestions are welcome...
Steve
Re: Very simple user query...
Posted by Michael Monnerie <m....@zmi.at>.
On Dienstag, 13. September 2005 17:38 Pedro Sam wrote:
> I haven't been using spamassassin for a while, but last I check,
> "spamassassin -r" will report spam to DCC/pyzor/razor all in one go.
Ah, so it's simple. I already have a script which takes from my SPAM_yes
folder and reports as SPAM, and from my SPAM_no folder to report a
false positive (aka HAM).
> Unfortunately, on its own, it doesn't address the user interface issue
> from the perspective of a client remotely accessing mail over
> IMAP/SMTP...
Here is how I do it:
# If you want to deliver the e-mail after learning back to cyrus:
fetchmail -a -s -n -p IMAP -u $user --folder 'SPAM_no' --auth 'password'
-m 'bash -c "/usr/bin/tee >(/usr/bin/sa-learn --ham --single
&>/dev/null)|/usr/bin/spamc|/usr/lib/cyrus-imapd/deliver"'
imap.host.domain
# This is the way I do it on normal imapd servers. That e-mail is learnt
and then discarded:
sudo -H -u $user fetchmail -a -s -n -p IMAP --folder 'SPAM_yes' --auth
'password' -m "bash -c \"tee >$checkfile|sa-learn --spam --single
&>/dev/null ; cat $headfile $checkfile >>$spamoutput.$user\""
imap.host.domain
Around this there's a loop with each user for who filtering is done. I
guess I should replace the "sa-learn" with "spamassassin -r" to report
to bayes/dcc/pyzor/razor, or does only sa-learn tell to bayes? Then I
should do both.
mfg zmi
--
// Michael Monnerie, Ing.BSc --- it-management Michael Monnerie
// http://zmi.at Tel: 0660/4156531 Linux 2.6.11
// PGP Key: "lynx -source http://zmi.at/zmi2.asc | gpg --import"
// Fingerprint: EB93 ED8A 1DCD BB6C F952 F7F4 3911 B933 7054 5879
// Keyserver: www.keyserver.net Key-ID: 0x70545879
Re: Very simple user query...
Posted by "Steve [Spamassasin]" <sp...@shic.co.uk>.
jdow wrote:
>> I absolutely do not want to report automatically - in the sense that
>> I am adamant that I want human intervention before reporting.
>> Conversely - given the task of establishing a remote shell; finding
>> the correct email in maildir - and verifying it is indeed the mail I
>> determined was a spam in my email client - followed by manually
>> reporting it individually to each service... I'm inclined not to
>> bother. If, for example I had an IMAP folder into which I drop spam
>> that my mail server should report on my behalf -then reporting would
>> become far less of a chore.
>
>
>
> Simple matter of coding. That is how I handle ham and spam training. I
> simply
> dunk it into ham and spam folders and let a cron job run sa-learn over
> the
> two folders. In this case you'd probably have to code up something
> that takes
> the folder apart properly, forwards the mail appropriately, then
> tosses it.
> I haven't done such a thing. But there are perl tools for reading
> messages
> via IMAP that could be used as the core of a new tool.
Hmmm - given that this seems such an obvious thing to want, and because
I'm quite laz^H^H^Hbusy these days, I'd hoped that there such thing
pre-existed. It strikes me that the best way to do this would be with a
daemon which monitors the IMAP folders for user-identified spam; salearn
and report it - then move it to the same folder as the automatically
identified spam. I realise that it wouldn't be a herculean effort to
implement this but I'm very reluctant to re-invent the wheel.
Re: Very simple user query...
Posted by jdow <jd...@earthlink.net>.
From: "Steve [Spamassasin]" <sp...@shic.co.uk>
> jdow wrote:
>
>> You do not say which version of spamassassin you are using. If it is not
>> 3.04 an upgrade might help.
>
> It's 3.04 - the latest stable build that's made it into "Gentoo Portage"
>
>>> * Is there somewhere where I can report spams which aren't caught by
>>> the default configuration in order to feed-back into future
>>> improvements?
>>
>> There are places to report them manually.
>
> I'm familiar with razor-report, for example - but it is a real pain to
> mess about with this command line tool when all my mail is managed
> remotely over IMAP
>
>> I have a strong personal bias against automating anything related to
>> spam REPORTING. Please examine the downsides of automatic reporting
>> before proceeding.
>
> I absolutely do not want to report automatically - in the sense that I am
> adamant that I want human intervention before reporting. Conversely -
> given the task of establishing a remote shell; finding the correct email
> in maildir - and verifying it is indeed the mail I determined was a spam
> in my email client - followed by manually reporting it individually to
> each service... I'm inclined not to bother. If, for example I had an IMAP
> folder into which I drop spam that my mail server should report on my
> behalf -then reporting would become far less of a chore.
Simple matter of coding. That is how I handle ham and spam training. I
simply
dunk it into ham and spam folders and let a cron job run sa-learn over the
two folders. In this case you'd probably have to code up something that
takes
the folder apart properly, forwards the mail appropriately, then tosses it.
I haven't done such a thing. But there are perl tools for reading messages
via IMAP that could be used as the core of a new tool.
{^_^}
Re: Very simple user query...
Posted by "Steve [Spamassasin]" <sp...@shic.co.uk>.
Pedro Sam wrote:
> > I'm familiar with razor-report, for example - but it is a real pain to
> > mess about with this command line tool when all my mail is managed
> > remotely over IMAP
>
> I haven't been using spamassassin for a while, but last I check,
> "spamassassin -r" will report spam to DCC/pyzor/razor all in one go.
(Having just checked the manual) - yes - that does seem much better than
reporting to all the services I use individually. Unfortunately, on its
own, it doesn't address the user interface issue from the perspective of
a client remotely accessing mail over IMAP/SMTP...
Re: Very simple user query...
Posted by Pedro Sam <pe...@tvworks.com>.
Steve [Spamassasin] wrote:
> I'm familiar with razor-report, for example - but it is a real pain to
> mess about with this command line tool when all my mail is managed
> remotely over IMAP
I haven't been using spamassassin for a while, but last I check,
"spamassassin -r" will report spam to DCC/pyzor/razor all in one go.
--
The authors know of one compiler that was written using only seven comments,
one of which read "This code is cursed."
page 731, The Dragon Book
Re: Very simple user query...
Posted by Rob Skedgell <ro...@nephelococcygia.demon.co.uk>.
On Wednesday 14 Sep 2005 22:44, jdow wrote:
> From: "Rob Skedgell" <ro...@nephelococcygia.demon.co.uk>
>
> On Tuesday 13 Sep 2005 21:15, Markus Eskola wrote:
> [...]
>
> > Just a quick question regarding the reporting... Do you guys report
> > all spam (including the once that SA allready caught) or only the
> > ones that got thru the net?
> >
> > Currently in my setup I have 3-4 diffrent users who move all the
> > spam that got thru into certain folders eg SPAM under IMAP. These
> > folders are scanned, emptied and reported once a night thru a
> > script. If someone has a more effectie way, I'd appreciate a hint
> > in the right direction.
>
> Most of it (5.0 <= score <= 30.0) gets LARTed by a java program that
> goes through the "confirmed spam" IMAP folder to the
^^^^^^^^^ e.g. *manually* confirmed as spam, not
just scored/flagged as such
[...]
> Ah, you are one of the people polluting the BLs. Thanks.... not.
No.
It was entirely my fault for not making it clearer that I do check the
confirmed spam folder very carefully first, before running the
reporting tool. It most certainly doesn't do anything like running from
cron, nor will it ever do that. If the IMAP seen flag isn't set on a
mail in that folder, it gets skipped as a safeguard against
carelessness on my part - the last thing I want is a mail that's just
been delivered to be reported without checking.
>
> Why not be a little saner and adopt a score higher than 5.0, a very
> marginal spam score, for reporting. That way you are not reporting
> false alarms and injuring innocent people.
See above. It's actually (score>=5.0 && manually_confirmed_as_spam)
I should stress that any mails I report are checked manually *first*.
False positives do *not* go to NANAS, SpamCop, the originating ISP etc.
False positives get dragged out of the spam folder, my whitelists fixed
(sometimes via whitelist_from_rcvd, sometimes in the PostgreSQL
database used by a couple of ACLs, depending on the context).
You can check the NANAS posts here
<http://groups.google.co.uk/groups?q=group:news.admin.net-abuse.sightings+nanas@nephelococcygia.demon.co.uk&start=0&scoring=d&>
if you like. See many false positives? No, nor me.
I very rarely mis-identify a false positive as spam, and on those rare
occasions the abuse contact who just got the LART in error gets a
grovelling apology from me for wasting their time.
--
Rob Skedgell <ro...@nephelococcygia.demon.co.uk>
Re: Very simple user query...
Posted by jdow <jd...@earthlink.net>.
From: "Rob Skedgell" <ro...@nephelococcygia.demon.co.uk>
On Tuesday 13 Sep 2005 21:15, Markus Eskola wrote:
[...]
> Just a quick question regarding the reporting... Do you guys report
> all spam (including the once that SA allready caught) or only the
> ones that got thru the net?
>
> Currently in my setup I have 3-4 diffrent users who move all the spam
> that got thru into certain folders eg SPAM under IMAP. These folders
> are scanned, emptied and reported once a night thru a script.
> If someone has a more effectie way, I'd appreciate a hint in the right
> direction.
Most of it (5.0 <= score <= 30.0) gets LARTed by a java program that
goes through the "confirmed spam" IMAP folder to the contacts.abuse.net
addresses for the IP address that sent to my MX, SpamCop and is also
posted to NANAS. If it scores over 30 it hits a discard ACL in exim.
Anything that sneaks through under 5.0 or went to a role account is also
singled out for extra vindictiveness and LARTed manually to anything
SpamTool missed and whois data checked very carefully for RFCI whois
eligibility (and a WDPRS report).
Oh, and I have a patched Mail::SpamAssassin::Plugin::URIDNSBL to pass
the domain names scanned over UDP to another listening application that
tests for "missing" entries in RFCI bogusmx and automatically sends the
submission by email. It also sends BCCs to postmaster@ and abuse@ so
that victims of "friendly fire" (through inadvertently using a CNAME
for their MX rather then deliberately registering 127.0.0.1) can get
unlisted.
++++++
Ah, you are one of the people polluting the BLs. Thanks.... not.
Why not be a little saner and adopt a score higher than 5.0, a very
marginal spam score, for reporting. That way you are not reporting
false alarms and injuring innocent people.
{^_^}
Re: Very simple user query...
Posted by Rob Skedgell <ro...@nephelococcygia.demon.co.uk>.
On Tuesday 13 Sep 2005 21:15, Markus Eskola wrote:
[...]
> Just a quick question regarding the reporting... Do you guys report
> all spam (including the once that SA allready caught) or only the
> ones that got thru the net?
>
> Currently in my setup I have 3-4 diffrent users who move all the spam
> that got thru into certain folders eg SPAM under IMAP. These folders
> are scanned, emptied and reported once a night thru a script.
> If someone has a more effectie way, I'd appreciate a hint in the right
> direction.
Most of it (5.0 <= score <= 30.0) gets LARTed by a java program that
goes through the "confirmed spam" IMAP folder to the contacts.abuse.net
addresses for the IP address that sent to my MX, SpamCop and is also
posted to NANAS. If it scores over 30 it hits a discard ACL in exim.
Anything that sneaks through under 5.0 or went to a role account is also
singled out for extra vindictiveness and LARTed manually to anything
SpamTool missed and whois data checked very carefully for RFCI whois
eligibility (and a WDPRS report).
Oh, and I have a patched Mail::SpamAssassin::Plugin::URIDNSBL to pass
the domain names scanned over UDP to another listening application that
tests for "missing" entries in RFCI bogusmx and automatically sends the
submission by email. It also sends BCCs to postmaster@ and abuse@ so
that victims of "friendly fire" (through inadvertently using a CNAME
for their MX rather then deliberately registering 127.0.0.1) can get
unlisted.
--
Rob Skedgell <ro...@nephelococcygia.demon.co.uk>
Re: Very simple user query...
Posted by Lefteris Tsintjelis <le...@ene.asda.gr>.
> Anybody got an idea
> how to prevent that confirmation?
Use spamcop_to_address "quick." instead of "submit." but thats something
you have to activate. The site has further info about this.
Re: Very simple user query...
Posted by Michael Monnerie <m....@zmi.at>.
On Mittwoch, 14. September 2005 16:12 Lefteris Tsintjelis wrote:
> Did I also mention the use of quite a few SPAM
> traps and grey listing (both are very effective).
Oh I love those, too *beg*
> Only if you are a registered (paid) user, then it is definetly worth
> reporting and things are listed relativly fast (I have a few
> objections to the exceptions he is making in favor of a large and
> pretty well known site, SPAM is SPAM no matter where it comes from)
> but I guess overall, its as you say it is. If you are not a
> registered user though IMHO then its a waste of resources.
I registered, but do not pay. I just changed my script to use
"spamassin" and not "sa-learn", now it reports to spamcop too. The
problem is, I get a mail per reported mail, where I have to click on a
link and press "confirm" on that page - annoying. Anybody got an idea
how to prevent that confirmation?
mfg zmi
--
// Michael Monnerie, Ing.BSc --- it-management Michael Monnerie
// http://zmi.at Tel: 0660/4156531 Linux 2.6.11
// PGP Key: "lynx -source http://zmi.at/zmi2.asc | gpg --import"
// Fingerprint: EB93 ED8A 1DCD BB6C F952 F7F4 3911 B933 7054 5879
// Keyserver: www.keyserver.net Key-ID: 0x70545879
Re: Very simple user query...
Posted by Lefteris Tsintjelis <le...@ene.asda.gr>.
>>I prefer to send it immediately which makes the updates of DCC and
>>razor even faster.
>
> How do you do it? Do you report back automatically every detected SPAM?
> That shouldn't be done, as I read from the homepage.
Not out of the box, I agree with that. I am using 3 threshold levels
and tested, trained and fined tuned the whole system for a while
before I turn on the auto reporting. Everything above a level, is auto
reported with a hit rate of 99.99%. I use a dedicated machine to
redirect, report and hold that SPAM for a while for this job only.
Everything in the middle I pass it through a couple of scripts,
analyze it, and what is left of it (not really much) manually report
it or take action against it to not enter the site again, but that
depends on the case. Did I also mention the use of quite a few SPAM
traps and grey listing (both are very effective).
>>What I am not so sure of is the SpamCop reporting.
>>It seems that its a complete waste since the black list that
>>maintains is not getting updated by any of those reports.
>
> AFAIK, spamcop sends e-mail to the admins responsible for that IP, and
> so it should help that ISPs get reports of zombies, relays, and so on.
> It fights on another level, but that one should be quite effective.
Only if you are a registered (paid) user, then it is definetly worth
reporting and things are listed relativly fast (I have a few
objections to the exceptions he is making in favor of a large and
pretty well known site, SPAM is SPAM no matter where it comes from)
but I guess overall, its as you say it is. If you are not a
registered user though IMHO then its a waste of resources.
Re: Very simple user query...
Posted by Michael Monnerie <m....@zmi.at>.
On Mittwoch, 14. September 2005 14:40 Lefteris Tsintjelis wrote:
> I prefer to send it immediately which makes the updates of DCC and
> razor even faster.
How do you do it? Do you report back automatically every detected SPAM?
That shouldn't be done, as I read from the homepage.
> What I am not so sure of is the SpamCop reporting.
> It seems that its a complete waste since the black list that
> maintains is not getting updated by any of those reports.
AFAIK, spamcop sends e-mail to the admins responsible for that IP, and
so it should help that ISPs get reports of zombies, relays, and so on.
It fights on another level, but that one should be quite effective.
mfg zmi
--
// Michael Monnerie, Ing.BSc --- it-management Michael Monnerie
// http://zmi.at Tel: 0660/4156531 Linux 2.6.11
// PGP Key: "lynx -source http://zmi.at/zmi2.asc | gpg --import"
// Fingerprint: EB93 ED8A 1DCD BB6C F952 F7F4 3911 B933 7054 5879
// Keyserver: www.keyserver.net Key-ID: 0x70545879
Re: Very simple user query...
Posted by Lefteris Tsintjelis <le...@ene.asda.gr>.
Michael Monnerie wrote:
> On Dienstag, 13. September 2005 22:15 Markus Eskola wrote:
>
>>Just a quick question regarding the reporting... Do you guys report
>>all spam (including the once that SA allready caught) or only the
>>ones that got thru the net?
All, with no exceptions made.
> I believe it should be done at least once per hour - so DCC and razor
> have it quickly detected. Otherwise, spammers have time until the night
> to send to a lot of servers. I currently do it in 10 minute intervals,
> as it doesn't really create too much load.
I prefer to send it immediately which makes the updates of DCC and
razor even faster. What I am not so sure of is the SpamCop reporting.
It seems that its a complete waste since the black list that maintains
is not getting updated by any of those reports.
Re: Very simple user query...
Posted by Michael Monnerie <m....@zmi.at>.
On Dienstag, 13. September 2005 22:15 Markus Eskola wrote:
> Just a quick question regarding the reporting... Do you guys report
> all spam (including the once that SA allready caught) or only the
> ones that got thru the net?
All - because others may have other rules, probably not identifying this
as SPAM. Imagine you get >5 points because your bayes is 100% sure, but
there's no hit on DCC, razor, etc. It's good for the others to report
it, so DCC and razor know it's SPAM, and therefore the next one who
receives it knows for sure about it.
> Currently in my setup I have 3-4 diffrent users who move all the spam
> that got thru into certain folders eg SPAM under IMAP. These folders
> are scanned, emptied and reported once a night thru a script.
> If someone has a more effectie way, I'd appreciate a hint in the
> right direction.
I believe it should be done at least once per hour - so DCC and razor
have it quickly detected. Otherwise, spammers have time until the night
to send to a lot of servers. I currently do it in 10 minute intervals,
as it doesn't really create too much load.
mfg zmi
--
// Michael Monnerie, Ing.BSc --- it-management Michael Monnerie
// http://zmi.at Tel: 0660/4156531 Linux 2.6.11
// PGP Key: "lynx -source http://zmi.at/zmi2.asc | gpg --import"
// Fingerprint: EB93 ED8A 1DCD BB6C F952 F7F4 3911 B933 7054 5879
// Keyserver: www.keyserver.net Key-ID: 0x70545879
Re: Very simple user query...
Posted by Stuart Johnston <st...@ebby.com>.
Michael Monnerie wrote:
> On Dienstag, 13. September 2005 14:06 Steve [Spamassasin] wrote:
>
> BTW: is pyzor good / worth the effort? It's latest release is September
> 7, 2002, for that I thought it wouldn't be used too much anymore. Do
> you get enough hits?
On my system, pyzor gets about half as many hits as DCC but is still one
of the more productive rules, coming in just after the SURBLs. However,
its default score is fairly high and it has a tendancy to FP on certain
types of messages so you have to be careful.
Here is a meta rule I was working on to reduce pyzor FPs:
meta L_pyzor_fp (PYZOR_CHECK && (MIME_HTML_ONLY || MIME_HTML_MOSTLY) &&
!DIGEST_MULTIPLE)
score L_pyzor_fp -2
Re: Very simple user query...
Posted by Markus Eskola <li...@eskola.nu>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Michael Monnerie wrote:
> On Dienstag, 13. September 2005 14:06 Steve [Spamassasin] wrote:
>
>> If, for example I had an IMAP folder into which I drop spam that
>> my mail server should report on my behalf -then reporting would
>> become far less of a chore.
>
>
> I'd be interested in this very much. Currently, I move SPAM into a
> folder, check that manually, and daily I move manually that
> confirmed SPAM into a special IMAP folder. That folder is scanned
> every 10 minutes, and it's content is learned by sa-learn and the
> e-mails deleted afterwards.
>
> Maybe it would be good to report that e-mails to razor, etc. too.
> I'll give it a try. Do you have a script to report from IMAP to SA?
>
>
> BTW: is pyzor good / worth the effort? It's latest release is
> September 7, 2002, for that I thought it wouldn't be used too much
> anymore. Do you get enough hits?
>
> mfg zmi
Just a quick question regarding the reporting... Do you guys report
all spam (including the once that SA allready caught) or only the ones
that got thru the net?
Currently in my setup I have 3-4 diffrent users who move all the spam
that got thru into certain folders eg SPAM under IMAP. These folders
are scanned, emptied and reported once a night thru a script.
If someone has a more effectie way, I'd appreciate a hint in the right
direction.
Sorry for hijacking ths thread a bit... ;)
/markus
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (MingW32)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFDJzNNY+EUV64Bqn0RAvmHAKDHfcy+mJ6BuR9GpbR5z/PsMZesrwCgidB2
gXf7TQ5sh/wEW730yPPFoTM=
=3kbx
-----END PGP SIGNATURE-----
Re: Very simple user query...
Posted by "Steve [Spamassasin]" <sp...@shic.co.uk>.
Michael Monnerie wrote:
> Maybe it would be good to report that e-mails to razor, etc. too. I'll
> give it a try. Do you have a script to report from IMAP to SA?
>
>
I don't... It can't be that hard to do using a polling approach... It
would be neater if this was triggered by the IMAP server... but I'm not
aware of such a facility. I'd love to be proved wrong...
> BTW: is pyzor good / worth the effort? It's latest release is
> September 7, 2002, for that I thought it wouldn't be used too much
> anymore. Do you get enough hits?
>
>
That's a good question... I wasn't aware that the latest release of
pyzor was so old... but it wouldn't have concerned me if I had... I'd be
inclined to suspect that what's important about pyzor is server-side.
Anyway - I've compiled some statistics over the past few months... and
it seems my installation of pyzor was really useful until sometime
during July... thereafter no more matches were made... which is curious...
#PYZOR_CHECK #spams
May 2794 3121
June 4402 4809
July 271 3017
August 0 3669
Sept 0 1546
This, I guess, might indicate part of the reason why less spam is caught
today... On further investigation I found pyzor crashed when run...
un-merging then re-merging it solved the problem which was probably some
strange python dependency.
Steve
Re: Very simple user query...
Posted by Michael Monnerie <m....@zmi.at>.
On Dienstag, 13. September 2005 14:06 Steve [Spamassasin] wrote:
> If, for
> example I had an IMAP folder into which I drop spam that my mail
> server should report on my behalf -then reporting would become far
> less of a chore.
I'd be interested in this very much. Currently, I move SPAM into a
folder, check that manually, and daily I move manually that confirmed
SPAM into a special IMAP folder. That folder is scanned every 10
minutes, and it's content is learned by sa-learn and the e-mails
deleted afterwards.
Maybe it would be good to report that e-mails to razor, etc. too. I'll
give it a try. Do you have a script to report from IMAP to SA?
BTW: is pyzor good / worth the effort? It's latest release is September
7, 2002, for that I thought it wouldn't be used too much anymore. Do
you get enough hits?
mfg zmi
--
// Michael Monnerie, Ing.BSc --- it-management Michael Monnerie
// http://zmi.at Tel: 0660/4156531 Linux 2.6.11
// PGP Key: "lynx -source http://zmi.at/zmi2.asc | gpg --import"
// Fingerprint: EB93 ED8A 1DCD BB6C F952 F7F4 3911 B933 7054 5879
// Keyserver: www.keyserver.net Key-ID: 0x70545879
Re: Very simple user query...
Posted by "Steve [Spamassasin]" <sp...@shic.co.uk>.
jdow wrote:
> You do not say which version of spamassassin you are using. If it is not
> 3.04 an upgrade might help.
It's 3.04 - the latest stable build that's made it into "Gentoo Portage"
>> * Is there somewhere where I can report spams which aren't caught by
>> the default configuration in order to feed-back into future
>> improvements?
>
> There are places to report them manually.
I'm familiar with razor-report, for example - but it is a real pain to
mess about with this command line tool when all my mail is managed
remotely over IMAP
> I have a strong personal bias against automating anything related to
> spam REPORTING. Please examine the downsides of automatic reporting
> before proceeding.
I absolutely do not want to report automatically - in the sense that I
am adamant that I want human intervention before reporting. Conversely
- given the task of establishing a remote shell; finding the correct
email in maildir - and verifying it is indeed the mail I determined was
a spam in my email client - followed by manually reporting it
individually to each service... I'm inclined not to bother. If, for
example I had an IMAP folder into which I drop spam that my mail server
should report on my behalf -then reporting would become far less of a chore.
Steve
Re: Very simple user query...
Posted by jdow <jd...@earthlink.net>.
From: "Steve [Spamassasin]" <sp...@shic.co.uk>
> I'm using spamassassin (Razor, Pyzor, DCC) and procmail to filter all my
> mail on my (Gentoo) linux-server, to which I connect from a number of
> Windows (XP/2000) machines using Mozilla Thunderbird to access my
> (dovecot) IMAP folders on the linux server. I configured spamassassin
> to use "Rulesdujour" and to regularly update those rules - and I was
> very happy... at least 99.99% of spam was correctly marked with only one
> incident of false positives (for which spamassasin wasn't entirely to
> blame.) in several months.
You do not say which version of spamassassin you are using. If it is not
3.04 an upgrade might help.
> Lately I've been less lucky - only ~99% of my spam is marked as such...
> which sounds good but the remaining 1% gives me up-to a dozen bogus
> messages each day... which is frustrating. To the naked eye the missed
> spam is obviously spam - but typically the only significant rule it
> triggers is the Bayesian rule... As I've stuck to the default settings
> this alone is insufficient to identify a mail as spam.
So far Bayes 99 triggers ONLY on spam here. I use a per user Bayes. I do
not use autolearn (or autowhitelist) and I basically train only with
caught spam that is not up to Bayes 99 or with spam that escaped. It has
taken some time; but, I've made it to one in one thousand spams or less
escaping detection with about the same level of false alarms.
> I'm left with several questions...
>
> * Is there somewhere where I can report spams which aren't caught by
> the default configuration in order to feed-back into future
> improvements?
There are places to report them manually.
> * Is there an easy way to report spam explicitly to the checksum
> services (Razor/Pyzor/DCC)?
I have a strong personal bias against automating anything related to
spam REPORTING. Please examine the downsides of automatic reporting
before proceeding.
{^_^}
Re: Very simple user query...
Posted by Michael Monnerie <m....@zmi.at>.
On Montag, 12. September 2005 11:27 Steve [Spamassasin] wrote:
> Lately I've been less lucky - only ~99% of my spam is marked as
> such...
Same for me: Getting some russian SPAM, which even is only sometimes
recognised by bayes. Could it be problems with cyrillic?
But I even get english SPAM that doesn't trigger. Even worse, it sayes
"bayes_00" and gives -2.599 points, effectively marking it as HAM...
mfg zmi
--
// Michael Monnerie, Ing.BSc --- it-management Michael Monnerie
// http://zmi.at Tel: 0660/4156531 Linux 2.6.11
// PGP Key: "lynx -source http://zmi.at/zmi2.asc | gpg --import"
// Fingerprint: EB93 ED8A 1DCD BB6C F952 F7F4 3911 B933 7054 5879
// Keyserver: www.keyserver.net Key-ID: 0x70545879
RE: Very simple user query...
Posted by Martin Hepworth <ma...@solid-state-logic.com>.
Steve
What version of SA and what URI-RBL's are you using??
--
Martin Hepworth
Snr Systems Administrator
Solid State Logic
Tel: +44 (0)1865 842300
-----Original Message-----
From: Steve [Spamassasin] [mailto:spamassassin_steve@shic.co.uk]
Sent: 12 September 2005 10:27
To: users@spamassassin.apache.org
Subject: Very simple user query...
I'm using spamassassin (Razor, Pyzor, DCC) and procmail to filter all my
mail on my (Gentoo) linux-server, to which I connect from a number of
Windows (XP/2000) machines using Mozilla Thunderbird to access my
(dovecot) IMAP folders on the linux server. I configured spamassassin
to use "Rulesdujour" and to regularly update those rules - and I was
very happy... at least 99.99% of spam was correctly marked with only one
incident of false positives (for which spamassasin wasn't entirely to
blame.) in several months.
Lately I've been less lucky - only ~99% of my spam is marked as such...
which sounds good but the remaining 1% gives me up-to a dozen bogus
messages each day... which is frustrating. To the naked eye the missed
spam is obviously spam - but typically the only significant rule it
triggers is the Bayesian rule... As I've stuck to the default settings
this alone is insufficient to identify a mail as spam.
I'm left with several questions...
* Is there somewhere where I can report spams which aren't caught by
the default configuration in order to feed-back into future
improvements?
* Is there an easy way to report spam explicitly to the checksum
services (Razor/Pyzor/DCC)?
Any other suggestions are welcome...
Steve
**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.
This footnote confirms that this email message has been swept
for the presence of computer viruses and is believed to be clean.
**********************************************************************