[Greylist-users] Database performance

[Brian Ross]

Some stats on our site (all numbers approximate):

13,000 mailboxes
21,000 total addressable user aliases (excluding mailing lists)

300,000 active triples in the greylist db over any given 24 hour 
period (last record update)
280,000 triples expire from the greylist db over any given 24 hour period
1,500,000 successful deliveries/day

Lots more connection attempts rejected by TCPwrappers and various 
throttling limits in sendmail

Over the past 15 days we're seeing average loads of 489 Kbps of port 
25 traffic inbound on our Internet border.  During the worm outbreaks 
in November and December we were averaging 1.14 Mbps, peaking at 
about 3.2 Mbps worth of SMTP crossing our Internet border.

Hope this helps.

Brian Ross


At 01:31 AM 1/23/2006, you wrote:

>Hello,
>
>Brian Ross said his database has around 1.2 million rows. Wayne database has
>around 4 million rows.
>
>May I have an idea of what trafic level are you handling ? (number or users or
>number of connections or messages a day)
>
>Thanks
>
>Jose-Marcio
>
>Wayne Walker wrote:
> > Follow Evan's advice, but also add indexes to the tables.  This greatly
> > improved the maintenance script run time for me.
> >
> > BE WARNED!  Adding the indexes will slam your mysql server for a LONG
> > TIME (I had 4 million rows in each table (maintenance db hadn't been
> > run in a long time) and it took about 15 minutes PER INDEX).
> >
> > Once I added the indexes, it still took a long time to run, but the
> > system was usable during the maintenance db run (it was NOT usable
> > before (load would hit 30's and 40's).
> >
> >
> > On Wed, Jan 04, 2006 at 11:05:06PM -0800, Brian Ross wrote:
> >
> >>Hello,
> >>
> >>I was wondering if anyone has any performance tuning tips for the MySQL
> >>db as implemented in relaydelay.
> >>
> >>We're running the db_maintenance script regularly, expiring
> >>approximately 1.25 million rows on a nightly basis with the maintenace
> >>taking approximately 1.5 hours.
> >>
> >>During the maintenance we're turning off our mail relays because the
> >>maintenance is so disk intensive that it causes relaydelay to slow down
> >>to the point of hitting the milter time out values (yes - I suppose I
> >>could look at altering those in the sendmail.cf).
> >>
> >>Any thoughts on optimization would be appreciated.
> >>
> >>
> >>-Brian
> >>
> >>
>
>
>--
>  ---------------------------------------------------------------
>  Jose Marcio MARTINS DA CRUZ           Tel. :(33) 01.40.51.93.41
>  Ecole des Mines de Paris              http://j-chkmail.ensmp.fr
>  60, bd Saint Michel                http://www.ensmp.fr/~martins
>  75272 - PARIS CEDEX 06      mailto:Jose-Marcio.Martins at ensmp.fr
>_______________________________________________
>Greylist-users mailing list
>Greylist-users at lists.puremagic.com
>http://lists.puremagic.com/cgi-bin/mailman/listinfo/greylist-users