On Mon, 25 Oct 2004, Kenneth Porter wrote:
> Right. The issue is with installation-time tests, not rules. There's a bug in
> a low-level Perl module that's only exercised at "make test" time.
Actually, I'm fairly certain it's also exercised if you run spamd as root.
Because I am one of the peo
I've been setting up my new box, and wanted to run some tests through it to
verify functionality.
I ran a bare email through it, with just the text of the GTUBE string in it,
and it doesn't seem to trip the system at all. I note in 20_body_tests.cf
the following:
body GTUBE /"elided,
Omar Armas wrote:
When you say score files, do you mean bayes and/or awl or do you mean
user preferences ?
For what I see the LDAP support for SA only stores the basic of user
preferences.
U, ummm, yah! Of course! heh, silly me.
I can say that with 35K+ users, mysql works just fine for
> When you say score files, do you mean bayes and/or awl or do you mean
> user preferences ?
For what I see the LDAP support for SA only stores the basic of user
preferences.
>
> I can say that with 35K+ users, mysql works just fine for user preferences.
>
Good to know it, thanks.
Omar
Omar Armas wrote:
Hi, I have qmail-ldap+qmail-scanner with some thousands of users.
I want to know if the ldap or mysql support in spamassassin 3 is stable enough
to use in such production enviroments(yes, i know it says "beta").
Has anyone used ldap or mysql to store score files? How many users
Hi, I have qmail-ldap+qmail-scanner with some thousands of users.
I want to know if the ldap or mysql support in spamassassin 3 is stable enough
to use in such production enviroments(yes, i know it says "beta").
Has anyone used ldap or mysql to store score files? How many users? Has it
been stab
FYI, I have discovered that some of the patches on the kernel mailing list
can trigger enough of the SARE rules to be declared as spam. The worst case
is one patch that made it to 14.9. I've had spams off that list that scored
less.
{^_-}
- Original Message -
From: "Kevin W. Gagel" <[EMAIL
* Kristopher Austin <[EMAIL PROTECTED]>:
> Matt,
>
> We've used SA for over two years now with settings similar to others
> that have replied. You should be fine with a stock SA 3.0.1 install.
> We greylist (you'll need other programs to do that) between 3 and 10,
But if you know that it scores
Came in this morning to a mess cause by SA crashing repeatedly over the
weekend. I had two messages in my mail queue that caused SA to crash
rather violently on Windows. After trying to disable various things I
gave up, downloaded SA 3.0.1 and installed it on my test machine where I
verified that t
Matt,
We've used SA for over two years now with settings similar to others
that have replied. You should be fine with a stock SA 3.0.1 install.
We greylist (you'll need other programs to do that) between 3 and 10,
tag as spam at 5 and delete at 10. I've never had one complaint about a
lost email
I don't simply tag the mail. Anything that scores at 5 or higher is tagged for
the user to delete themselves if it arrives in their inbox. If it scores at 10
or higher I have the system delete it automatically - no one ever see's it.
I started to introduce it to them three years ago when our spam
Elsa Andrés wrote:
Wow... I'm impressed!
I added "-x" option to spamd start script as you mentioned (to prevent "user
prefs" to become active) and, sure it works!
Thank you very much for you tip. I was being crazy to find the error, and
looked and re-looked the local.cf and anything seemed O.K. I t
On Mon, 2004-10-25 at 09:01, Theo Van Dinter wrote:
> There is a list of dependencies in the INSTALL file, but we don't list
> dependencies which come standard with Perl.
Is there any way to find out exactly what the standard perl deps of SA
are? Right now we bundle perl inside one of our
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Jeff Chan writes:
> On Monday, October 25, 2004, 1:26:55 PM, Justin Mason wrote:
> > http://www.projecthoneypot.org/
>
> > seems interesting, they plan to share their resulting corpora, and they
> > seem like nice guys too, from what I recall from wh
On Fri, 22 Oct 2004 22:15:33 -0700 (PDT), "email builder"
<[EMAIL PROTECTED]> said:
> May I ask either of you how you installed SA? I am pretty much at a
> loss... just grabbing for whatever I can get at this point...
We installed Debian Woody and grabbed the Spamassassin 2.64 package from
www.b
On Monday, October 25, 2004, 1:26:55 PM, Justin Mason wrote:
> http://www.projecthoneypot.org/
> seems interesting, they plan to share their resulting corpora, and they
> seem like nice guys too, from what I recall from when we saw them at the
> MIT Spam Conference (IIRC) ;)
> --j.
I think these
Kevin, if you simply mark the mail as spam and do no further processing
after SpamAssassin there is no censorship involved. All you're doing is
telling the recipient that this mail is probably spam. With a little
experience the users will figure out how to sort their email so that
the egregious spa
Matt <[EMAIL PROTECTED]> writes:
> Does anyone have experience with using spamassassin and rule sets at a
> college environment? We'd like to block just mail that is one
> hundred percent spam, and not risk blocking false positives. Any
> thoughts or ideas?
You can approach 100% accuracy to bl
http://www.projecthoneypot.org/
seems interesting, they plan to share their resulting corpora, and they
seem like nice guys too, from what I recall from when we saw them at the
MIT Spam Conference (IIRC) ;)
--j.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Bill Landry writes:
> I've seen a couple of people ask about this now on the amavis list, and I
> was wonder what it was when I saw it, as well. What do the following
> "inhibited further callbacks" lines mean any why are they reported in debug
> out
On Mon, 2004-10-25 at 14:35 -0400, Matt wrote:
> Hi,
> Does anyone have experience with using spamassassin and rule sets at a
> college environment? We'd like to block just mail that is one
> hundred percent spam, and not risk blocking false positives. Any
> thoughts or ideas?
I'd recommend you
This is what we all strive for in any enviorment, not just college. I would
try the vanilla install of SA and SURBL. There are a few rulesets on the
SARE site that should work well with no FPs. But start with general install
first.
Also setup a test system, and grab a few people to test it out. Y
Matt,
See http://avas.cnc.bc.ca for what were doing here. We are only sitewide and
do not yet have individual settings available. But we are a college and all
email is scanned. I setup this site to answer questions that my users might
have and explain how things are setup. I figured informing them
I've seen a couple of people ask about this now on the amavis list, and I
was wonder what it was when I saw it, as well. What do the following
"inhibited further callbacks" lines mean any why are they reported in debug
output? URIDNSBL checks appear to be working fine, I was just wondering
what t
At 08:55 PM 10/24/2004, Tremaine wrote:
We recently implemented spamassassin on our server to mitigate spam
volumes, and it does it quite well... except for one thing. For
reasons I can't determine, it will suddenly start chewing through RAM
and cpu usage until either spamd dies or the server beco
Hi,
Does anyone have experience with using spamassassin and rule sets at a
college environment? We'd like to block just mail that is one
hundred percent spam, and not risk blocking false positives. Any
thoughts or ideas?
At 07:31 AM 10/23/2004, Kai Schaetzl wrote:
> - excessive memory-usage fixes
>
installed on two machines, no problems so far. Memory usage of MailScanner
is slightly higher after upgrade. I assume the memory-usage fixes were for
spamd, anyway?
Kai,
Looking in the Changes a couple of fixes, none see
--On Monday, October 25, 2004 12:56 PM -0400 "Christopher X. Candreva"
<[EMAIL PROTECTED]> wrote:
Thanks. So the DNSBL checks run, just not 'make test'.
Right. The issue is with installation-time tests, not rules. There's a bug
in a low-level Perl module that's only exercised at "make test" tim
Josh Trutwin wrote:
On Mon, 25 Oct 2004 12:40:44 -0400
Rick Macdougall <[EMAIL PROTECTED]> wrote:
Hi,
http://wiki.apache.org/spamassassin/WebUserInterfaces
I'll be updating the php-sa-mysql2 to work with 3.x sometime in the
next few days.
Out of curiousity, what needs to change from version 2.x
On Mon, 25 Oct 2004 12:40:44 -0400
Rick Macdougall <[EMAIL PROTECTED]> wrote:
> Hi,
>
> http://wiki.apache.org/spamassassin/WebUserInterfaces
>
> I'll be updating the php-sa-mysql2 to work with 3.x sometime in the
> next few days.
Out of curiousity, what needs to change from version 2.x to 3.x?
There's an old project called "WebUserPrefs" written in PHP for
modifying end-user preferences via the web. It's not applicable to
SA-3.0, and I wonder if there are other similar projects out there yet
for SA-3.0... ?
The SASQL plugin for SquirrelMail works just fine with SA-3.0
http://www.squirr
Hi, even using spamd with -c flag (spamd -d -c -v -u vpopmail
--pidfile=/home/vpopmail/etc/spamd.pif) sometimes it can't create the
user_prefs file.
The permissions are vpopmail.vchkpw for all domains. Sometimes it creates
user_prefs file without problem, and sometimes it gives this message:
On Mon, 25 Oct 2004, Justin Mason wrote:
> > Does this mean if spamd is running as root, it runs the tests as someone
> > other than root, or that it doesn't run the tests at all ?
>
> not at all.
Thanks. So the DNSBL checks run, just not 'make test'.
=
Forrest Aldrich wrote:
There's an old project called "WebUserPrefs" written in PHP for
modifying end-user preferences via the web. It's not applicable to
SA-3.0, and I wonder if there are other similar projects out there yet
for SA-3.0... ?
Hi,
http://wiki.apache.org/spamassassin/WebUserInterfac
There's an old project called "WebUserPrefs" written in PHP for
modifying end-user preferences via the web. It's not applicable to
SA-3.0, and I wonder if there are other similar projects out there yet
for SA-3.0... ?
Thanks.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Christopher X. Candreva writes:
> On Fri, 22 Oct 2004, Justin Mason wrote:
>
> > - avoid bug in Sys::Hostname::Long that renames the hostname when "make
> > test" is run
>
> According to changelog:
>
> bug 3806: do not run DNSBL and SPF tes
>Matt Kettler wrote:
>> Suggestion: do ONE whitelist per whitelist_from statement, not two.
>> whitelist_from should only accept one parameter.
>>
>> whitelist_from_rcvd expects two parameters, but the second parameter
>> is not a From: address, so perhaps looking at those got you
>> confused..
>
>
>You're probably getting hit by bug 3855 ('whitelist_from in local.cf
>ignored after whitelist_from loaded from users prefs').
>
>Either disable the user_prefs (if you are not using them) by giving
>spamd the '-x' option, or upgrade to 3.0.1, where it's fixed.
Hi Marco,
Wow... I'm impressed!
I
On Fri, 22 Oct 2004, Justin Mason wrote:
> - avoid bug in Sys::Hostname::Long that renames the hostname when "make
> test" is run
According to changelog:
bug 3806: do not run DNSBL and SPF tests as root on non-linux UNIX
platforms, due to a stupid bug in Sys::Hostname::Long that renam
On Mon, Oct 25, 2004 at 09:30:27AM +0200, Henri van Riel wrote:
> Is there a list of dependencies for SA so I only have to install what
> I need for SA to run?
There is a list of dependencies in the INSTALL file, but we don't list
dependencies which come standard with Perl.
--
Randomly Generated
Matt Kettler wrote:
> Suggestion: do ONE whitelist per whitelist_from statement, not two.
> whitelist_from should only accept one parameter.
>
> whitelist_from_rcvd expects two parameters, but the second parameter
> is not a From: address, so perhaps looking at those got you
> confused..
Barring
At 11:54 AM 10/25/2004 +0200, Elsa Andrés wrote:
whitelist_from [EMAIL PROTECTED] [EMAIL PROTECTED]
whitelist_from [EMAIL PROTECTED] [EMAIL PROTECTED]
whitelist_from [EMAIL PROTECTED] [EMAIL PROTECTED]
whitelist_from [EMAIL PROTECTED] [EMAIL PROTECTED]
But e-mails coming from "[EMAIL PROTECTED]" or
At 07:37 AM 10/25/2004 +, tadeusz ferenc wrote:
Using spamassassin 3.0.0 on my ISP whatever I set the score to in my rule
set, when the event is triggered the score is always 1. I am using the
following as a test
add_header all tf Score Values (_TESTSSCORES(,)_)
rawbody TF_RAWBODY_WARNING_1
I also would add Net::DNS, Mail::SPF::Query, and Razor. SA 3.0 has been
99% effective.
> Some dependencies i recall are
>
> perl-Time-HiRes_1.38-4_i386.rpm
> perl-Digest-HMAC_1.01-11_noarch.rpm
> perl-Digest-SHA1_2.07-1.rhfc1.dag_i386.rpm
> perl-Net-DNS_0.31-3.2_noarch.rpm
>
> Correct me if i am wr
i really dont think there is a need of rolex specific ruleset, Razor,
DCC and URI checks took care of them for me.
Peter Clark wrote:
Apparently hawking Rolexes is the in thing with spammers these days.
I haven't seen any rulesets around that would help combat it, so I
wrote one.
It's availa
Some dependencies i recall are
perl-Time-HiRes_1.38-4_i386.rpm
perl-Digest-HMAC_1.01-11_noarch.rpm
perl-Digest-SHA1_2.07-1.rhfc1.dag_i386.rpm
perl-Net-DNS_0.31-3.2_noarch.rpm
Correct me if i am wrong or have missed some thing.
Is there a list of dependencies for SA so I only have to install what
I
Elsa Andrés wrote:
Any idea on how to get around this issue will be appreciated.
You're probably getting hit by bug 3855 ('whitelist_from in local.cf
ignored after whitelist_from loaded from users prefs').
Either disable the user_prefs (if you are not using them) by giving
spamd the '-x' option,
Hello,
I am trying to get manual whitelisting to work without success.
Some background:
- SA is running as daemon every time the system is started
- SA version is 3.0.0
- SA is installed on a SuSE 8.2 box
- Using /etc/mail/spamassassin/local.cf sitewide rules
- E-mails trajectory is as follows:
Using spamassassin 3.0.0 on my ISP whatever I set the score to in my rule
set, when the event is triggered the score is always 1. I am using the
following as a test
add_header all tf Score Values (_TESTSSCORES(,)_)
rawbody TF_RAWBODY_WARNING_1 /message\.scr/i
score TF_RAWBODY_WARNING_1 20
As spa
Hello all,
I'm new to the list (and new to SpamAssassin as well), so hello to you
all!
I was wondering if I'm going to need a full install of Perl on the
machine I would like to run SA on? Perl is a bit big and I don't
like the idea of having a full set of Perl on my mailserver.
Is there a list
On Saturday, October 23, 2004, 4:36:50 AM, Sven Ehret wrote:
> I installaed a Postfix/SA Mailrelay for one of our clients and it's
> performing good, /except/, and this could be critical, there are no
> SBL, RBL, SURBL or similar checks made. This leads to inacceptable low
> scorings and could be a
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Sahil Tandon writes:
> Jeff Chan wrote:
>
> > On Sunday, October 24, 2004, 3:09:53 PM, John Andersen wrote:
> >
> >>What file are you finding this above bug in?
> >>I don't see that anywhere on my 3.0.1 install!
> >
> >
> > There should be some
Jeff Chan wrote:
On Sunday, October 24, 2004, 3:09:53 PM, John Andersen wrote:
>
What file are you finding this above bug in?
I don't see that anywhere on my 3.0.1 install!
There should be some kind of change log included in the
distribution.
Indeed there is; aptly labeled CHANGES.
--
Sahil Tand
Hi Peter,
You use the * character in body rules. For example, you do:
body __REAL_ROLEX_REPLICA1 /real.*replica.*role.?x/i
I got badly burned doing this sort of thing early on. These rules
consume huge amounts of memory and processor time. I would do
something like this instead:
body _
We recently implemented spamassassin on our server to mitigate spam
volumes, and it does it quite well... except for one thing. For
reasons I can't determine, it will suddenly start chewing through RAM
and cpu usage until either spamd dies or the server becomes
unresponsive.
There are a lot of en
On Sunday, October 24, 2004, 3:09:53 PM, John Andersen wrote:
> On Sunday 24 October 2004 01:16 pm, Burnie wrote:
>> "Bill Landry" <[EMAIL PROTECTED]> writes:
>> > urirhssub URIBL_JP_SURBL multi.surbl.org. A 64
>> > headerURIBL_JP_SURBL eval:check_uridnsbl('URIBL_JP_SURBL')
>>
>> ^^
>
56 matches
Mail list logo
