Hello Gary,

is a few mounths that I use amavisd-new/spamassassin with DB.

Really, It's happening that the data size become huge.. I have executed the
syncs/ force expire option for sa-learn.. I execute the dump-magic option
here the output before and after launching the commands:

av3:~ # sa-learn -u amavis --dump magic
0.000          0          3          0  non-token data: bayes db version
0.000          0     750713          0  non-token data: nspam
0.000          0    1762742          0  non-token data: nham
0.000          0     474911          0  non-token data: ntokens
0.000          0 1141033678          0  non-token data: oldest atime
0.000          0 1141310504          0  non-token data: newest atime
0.000          0          0          0  non-token data: last journal sync
atime
0.000          0 1141277737          0  non-token data: last expiry atime
0.000          0      43200          0  non-token data: last expire atime
delta
0.000          0     284156          0  non-token data: last expire
reduction count
av3:~ # sa-learn -u amavis --dump magic
av3:~ # su amavis -c 'sa-learn --sync --force-expire'
expired old bayes database entries in 29 seconds
358017 entries kept, 117564 deleted
token frequency: 1-occurrence tokens: 65.41%
token frequency: less than 8 occurrences: 20.88%
av3:~ # sa-learn -u amavis --dump magic
0.000          0          3          0  non-token data: bayes db version
0.000          0     750751          0  non-token data: nspam
0.000          0    1762839          0  non-token data: nham
0.000          0     360606          0  non-token data: ntokens
0.000          0 1141033678          0  non-token data: oldest atime
0.000          0 1141310986          0  non-token data: newest atime
0.000          0          0          0  non-token data: last journal sync
atime
0.000          0 1141310570          0  non-token data: last expiry atime
0.000          0      43200          0  non-token data: last expire atime
delta
0.000          0     117564          0  non-token data: last expire
reduction count

It's all OK or somethings does not work?

Moreover, I have seen that the ntokens is much more the number you said
below.. How I have to do?

Thanks,

rocsca

----- Original Message ----- 
From: "Gary V" <[EMAIL PROTECTED]>
To: <[email protected]>
Sent: Thursday, October 20, 2005 3:57 PM
Subject: Re: [AMaViS-user] About SA...


> Rocco wrote:
>
> > Yet I have not imported AWL... How I have to do to import it?
>
> If you like, you can dig through this document. You create the table
> using the supplied awl_mysql.sql file, something like
> "mysql -p sa_bayes < awl_mysql.sql" and then use the
> convert_awl_dbm_to_sql script to import the data, then make the
> necessary changes in local.cf. The script can be tricky, so make sure
> you understand the entire process before you try it.
> http://www200.pair.com/mecham/spam/debian-spamassassin-sql.html
>
> > I don't know if I syncing or expiring data in bayes_seen table on
regular
> > basis (the SA local.cf file is the original with the options for
accessing
> > DB).
>
> If you have not disabled it, then this happens automatically.
>
http://spamassassin.apache.org/full/3.1.x/dist/doc/sa-learn.html#expiration
>
> >  What is the difference between sync and expire? How I can setup the two
> > operations on regular basis? Should I reduce in this manner the
dimension of
> > the DB?
>
> It is my understanding Bayes stores recent data in one table (journal)
> and the main data in another. The sync operation updates the main
> database with the recent data. Expire attempts to get rid of old stale
> data.
>
> > PS: I can't clean the DB (Otherwise I'ld loose all the spam information
> > collected till now)
>
> The --sync --force-expire may clean up old tokens, but it does not
> loose all the spam information.
>
> > PPS: Can I add more spam information from a mail box (maildir format)
> > containing messages filtered with a spam scanner on a client side?
>
> http://spamassassin.apache.org/full/3.1.x/dist/doc/sa-learn.html#options
>
> > PPPS: Is important to import the user preference too in the bayes DB?
> > Thanks
> > rocsca
>
> I wanted you to run:
> sa-learn -u vscan --dump magic
> to see if your database is larger than it should be.
> Run it and look at the 'ntokens' number:
> 0.000          0     158445          0  non-token data: ntokens
> Here mine shows about 160,000 tokens which is normal. If yours shows a
> much larger number, it could indicate a problem.
>
> Also, I see you may be moving your data to InnoDB. From a previous
> post, it looks like you may not be running MySQL version 4.1. What
> version are you running? (mysql -V). If you are running older 3.xx
> versions of MySQL, you may need to make some changes to my.cnf.
> Versions 4.0 and newer do not require any change to my.cnf. If you are
> using 3.xx, see:
>
http://www.doctaur.com/dtdocs/databases/mysql-manual/innodb-configuration.html#SEC637
>
http://www.paulstimesink.com/index.php?op=ViewArticle&articleId=167&blogId=2
>
> Gary V
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> AMaViS-user mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/amavis-user
> AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
> AMaViS-HowTos:http://www.amavis.org/howto/



-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
_______________________________________________
AMaViS-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/

Reply via email to