You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Michael Parker <pa...@pobox.com> on 2004/10/08 18:51:47 UTC

Bayes/AWL SQL Roll Call

Hi All,

I'm compling some stats on Bayes and AWL in SQL for my upcoming talk
at ApacheCon[1].  If you're making use of either Bayes and/or AWL with
SQL based storage could you please respond (off list is fine) to this
email. I will only be using raw numbers, no names. In particular, I'm
interested in:

1) Bayes, AWL or both?

2) What DB server you are using.

3) Sitewide or Individual

4) Approximate mail traffic per day

5) Approximate number of users

6) Approximate DB size
  o physical size (ie on disk)
  o select count(*) from bayes_vars;
  o select count(*) from bayes_token;
  o select count(*) from awl;
  o select count(distinct(username)) from awl;

7) Multiple DB servers? If so, how many?

8) Multiple mail servers talking to a DB server/cluster? If so, how
   many?

9) Are you using the default config? or have you made local
   modifications?  If you've made local modifications, can you tell me
   a little bit about them?

9) Initial impressions

Thanks in advance for any input,
Michael Parker

1: http://www.apachecon.com/
  

Re: Bayes/AWL SQL Roll Call

Posted by Ryan Moore <ry...@perigee.net>.
Michael Parker wrote:
> Hi All,
> 
> I'm compling some stats on Bayes and AWL in SQL for my upcoming talk
> at ApacheCon[1].  If you're making use of either Bayes and/or AWL with
> SQL based storage could you please respond (off list is fine) to this
> email. I will only be using raw numbers, no names. In particular, I'm
> interested in:
> 
> 1) Bayes, AWL or both?

Just Bayes

> 2) What DB server you are using.

Mysql 4.0

> 3) Sitewide or Individual

Sitewide

> 4) Approximate mail traffic per day

20-30k messages per day

> 5) Approximate number of users

~3000

> 6) Approximate DB size
>   o physical size (ie on disk)
200MB

>   o select count(*) from bayes_vars;
2

>   o select count(*) from bayes_token;
565,221

bayes_seen is 1,550,911

> 7) Multiple DB servers? If so, how many?
Just one

> 8) Multiple mail servers talking to a DB server/cluster? If so, how
>    many?

Just a stand alone box

> 9) Are you using the default config? or have you made local
>    modifications?  If you've made local modifications, can you tell me
>    a little bit about them?

Using amavisd-new 2.1 in conjunction with SA3.0, with several of the 
rulesets from rulesemporium.com, a few custom rules, but with 
Bayes+SURBL+rulesemporium there really isn't anything that gets through.

> 9) Initial impressions
10? ;]
I love it. Since I upgraded to SA3 and amavisd 2 (was running sa2.6 and 
an older amavis), I can count the number of spams that made it through 
in the past three weeks on one hand.

> Thanks in advance for any input,
> Michael Parker
> 
> 1: http://www.apachecon.com/


Ryan Moore
----------
Perigee.net Corporation
704-849-8355 (sales)
704-849-8017 (tech)
www.perigee.net