Re[2]: (Upgrade 2.63 - 3.0.0) AND (Upgrade Bayes File - MySQL) Documentation?

2004-09-27 Thread Robert Menschel
Hello Kai,

Sunday, September 26, 2004, 2:17:37 PM, you wrote:

KS Robert Menschel wrote on Sat, 25 Sep 2004 14:00:07 -0700:

 We split them out this way specifically because file 0 gives the most
 bang for the buck (er, computer resource), file 1 also gives good return
 on the investment, and files 2 and 3 are useful for aggressive sites with
 lots of computer resources, but sites that are low on resources should
 probably not use these low-return files.


KS I was gradually aware of these split versions, but I didn't make the
KS connection from your answer to that, thanks. I use rules_du_jour and there
KS doesn't seem to be another option as to use f.i. sare_header in general
KS unless I make up my own rules. Or does it use the 0 rules only, anyway?

As far as I know, RDJ lets you use any individual file or combination of
files -- you just need to specify the correct parameters to feed RDJ, so
it knows what to pull in for you. I don't use it, and haven't looked at
it in any detail, so I can't give you more specific help, but others
should be able to.

Bob Menschel





Re: (Upgrade 2.63 - 3.0.0) AND (Upgrade Bayes File - MySQL) Documentation?

2004-09-26 Thread Kai Schaetzl
Robert Menschel wrote on Sat, 25 Sep 2004 14:00:07 -0700:

 We split them out this way specifically because file 0 gives the most
 bang for the buck (er, computer resource), file 1 also gives good return
 on the investment, and files 2 and 3 are useful for aggressive sites with
 lots of computer resources, but sites that are low on resources should
 probably not use these low-return files.


I was gradually aware of these split versions, but I didn't make the 
connection from your answer to that, thanks. I use rules_du_jour and there 
doesn't seem to be another option as to use f.i. sare_header in general 
unless I make up my own rules. Or does it use the 0 rules only, anyway?


Kai

-- 

Kai Schätzl, Berlin, Germany
Get your web at Conactive Internet Services: http://www.conactive.com
IE-Center: http://ie5.de  http://msie.winware.org





Re: (Upgrade 2.63 - 3.0.0) AND (Upgrade Bayes File - MySQL) Documentation?

2004-09-25 Thread Kai Schaetzl
Robert Menschel wrote on Wed, 22 Sep 2004 15:35:35 -0700:

 If they are, trim your 70_sare_header and 70_sare_html down to the
 file 0 file for best performance.


What do you mean by this?


Kai

-- 

Kai Schätzl, Berlin, Germany
Get your web at Conactive Internet Services: http://www.conactive.com
IE-Center: http://ie5.de  http://msie.winware.org





(Upgrade 2.63 - 3.0.0) AND (Upgrade Bayes File - MySQL) Documentation?

2004-09-22 Thread Klaus Pforte
Hi there,
I'm running 2.63 with sitewide bayes and whitelist in a normal file.
It runs good!
Now I want to upgrade to 3.0.0 final and want to use the new feature 
with Bayes and Whitelist(?) in a MySQL database.

But I can't find a good documentation to do this.
Or did I search not long enough...?
Another question:
I'm running 2.63 with Pyzor, DCC, Razor2 and the following rulesets:
70_sare_header
70_sare_html
70_sare_random
70_sare_spoof
99_sare_adult
99_sare_fraud
antidrug
bogus_virus_warnings
evilnumbers
spamcop_uri (0.22)
I have spamd processes with around 30 MB.
What rulesets in special make the most memory consumption?
And:
Need 3.0.0 more memory?
My server has 512MB, but need most of it for HTTP.
Thanks for help, greetings from Germany,
Klaus



Re: (Upgrade 2.63 - 3.0.0) AND (Upgrade Bayes File - MySQL) Documentation?

2004-09-22 Thread Michael Parker
On Wed, Sep 22, 2004 at 08:01:36PM +0200, Klaus Pforte wrote:
 
 I'm running 2.63 with sitewide bayes and whitelist in a normal file.
 It runs good!
 

Currently you can't do sitewide AWL in SQL.

 Now I want to upgrade to 3.0.0 final and want to use the new feature 
 with Bayes and Whitelist(?) in a MySQL database.
 
 But I can't find a good documentation to do this.
 Or did I search not long enough...?
 

sql/README*

 
 Another question:
 I'm running 2.63 with Pyzor, DCC, Razor2 and the following rulesets:
 70_sare_header
 70_sare_html
 70_sare_random
 70_sare_spoof
 99_sare_adult
 99_sare_fraud
 antidrug
 bogus_virus_warnings
 evilnumbers
 spamcop_uri (0.22)

For sure get rid of spamcop_uri, it's built into 3.0.0.

I would actually suggest, removing all of the above (rulesets, you can
keep Pyzor/DCC/Razor around if you'd like), start with a new install
then add only as needed.  Extra rulesets do add to your memory
requirements.

Michael


Re[2]: (Upgrade 2.63 - 3.0.0) AND (Upgrade Bayes File - MySQL) Documentation?

2004-09-22 Thread Robert Menschel
Hello Klaus,

Wednesday, September 22, 2004, 11:07:19 AM, Michael wrote:

MP On Wed, Sep 22, 2004 at 08:01:36PM +0200, Klaus Pforte wrote:
 I'm running 2.63 with sitewide bayes and whitelist in a normal file.
 It runs good!

 Another question:
 I'm running 2.63 with Pyzor, DCC, Razor2 and the following rulesets:
 70_sare_header
 70_sare_html
 70_sare_random
 70_sare_spoof
 99_sare_adult
 99_sare_fraud
 antidrug
 bogus_virus_warnings
 evilnumbers
 spamcop_uri (0.22)

MP For sure get rid of spamcop_uri, it's built into 3.0.0.

Ditto with antidrug.

MP I would actually suggest, removing all of the above (rulesets, you can
MP keep Pyzor/DCC/Razor around if you'd like), start with a new install
MP then add only as needed.  Extra rulesets do add to your memory
MP requirements.

99_sare_adult is obsolete, replaced by 70_sare_adult long time ago.

70_sare_adult itself has not been modified since June, and so might be
losing some potency. You might want to leave it out of your 3.0 setup
unless its target spam sneak through.

If resources are not a problem, add some of the 70_sare_genlsubj files.
If they are, trim your 70_sare_header and 70_sare_html down to the
file 0 file for best performance.

Bob Menschel