Re: Performance problems testing with Spamassassin 3.1.0

From: Andrew McMillan
Subject: Re: Performance problems testing with Spamassassin 3.1.0
Date: ,
Msg-id: 1122668371.5691.352.camel@lamb.mcmillan.net.nz
(view: Whole thread, Raw)
In response to: Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher)
Responses: Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher)
Re: Performance problems testing with Spamassassin 3.1.0  (Tom Lane)
List: pgsql-performance

Tree view

Performance problems testing with Spamassassin 3.1.0 Bayes module.  (Matthew Schumacher, )
 Re: Performance problems testing with Spamassassin 3.1.0 Bayes module.  (Josh Berkus, )
  Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
   Re: Performance problems testing with Spamassassin 3.1.0 Bayes module.  (Josh Berkus, )
    Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
   Re: Performance problems testing with Spamassassin 3.1.0  (Dennis Bjorklund, )
    Re: Performance problems testing with Spamassassin 3.1.0  (Josh Berkus, )
     Re: Performance problems testing with Spamassassin 3.1.0  (John Arbash Meinel, )
 Re: Performance problems testing with Spamassassin 3.1.0 Bayes module.  (Tom Lane, )
  Re: Performance problems testing with Spamassassin  ("Luke Lonergan", )
 Re: Performance problems testing with Spamassassin 3.1.0  (Karim Nassar, )
  Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
   Re: Performance problems testing with Spamassassin 3.1.0  (Gavin Sherry, )
    Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
     Re: Performance problems testing with Spamassassin 3.1.0  (Gavin Sherry, )
   Re: Performance problems testing with Spamassassin 3.1.0  (Andrew McMillan, )
    Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
     Re: Performance problems testing with Spamassassin 3.1.0  (PFC, )
     Re: Performance problems testing with Spamassassin 3.1.0  (Andrew McMillan, )
      Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
      Re: Performance problems testing with Spamassassin 3.1.0  (Tom Lane, )
       Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
        Re: Performance problems testing with Spamassassin 3.1.0  (Tom Lane, )
         Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
          Re: Performance problems testing with Spamassassin 3.1.0  (John Arbash Meinel, )
         Re: Performance problems testing with Spamassassin 3.1.0  (Michael Parker, )
          Re: Performance problems testing with Spamassassin 3.1.0  (Tom Lane, )
          Re: Performance problems testing with Spamassassin 3.1.0  (Tom Lane, )
           Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
         Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
          Re: Performance problems testing with Spamassassin 3.1.0  (PFC, )
           Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
            Re: Performance problems testing with Spamassassin 3.1.0  ("Jim C. Nasby", )
             Re: Performance problems testing with Spamassassin 3.1.0  (Michael Parker, )
              Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
               Re: Performance problems testing with Spamassassin 3.1.0  (PFC, )
               Re: Performance problems testing with Spamassassin 3.1.0  (John A Meinel, )
                Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
               Re: Performance problems testing with Spamassassin 3.1.0  (Tom Lane, )
                Re: Performance problems testing with Spamassassin 3.1.0  (John A Meinel, )
                 Re: Performance problems testing with Spamassassin 3.1.0  (Tom Lane, )
                Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
                 Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
                  Re: Performance problems testing with Spamassassin 3.1.0  (John A Meinel, )
                   Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
                    Re: Performance problems testing with Spamassassin 3.1.0  (John A Meinel, )
        Re: Performance problems testing with Spamassassin 3.1.0  (John Arbash Meinel, )
         Re: Performance problems testing with Spamassassin 3.1.0  (John Arbash Meinel, )
          Re: Performance problems testing with Spamassassin 3.1.0  (Matthew Schumacher, )
           Re: Performance problems testing with Spamassassin 3.1.0  ("Jim C. Nasby", )
            Re: Performance problems testing with Spamassassin 3.1.0  (Andreas Pflug, )
 Re: Performance problems testing with Spamassassin 3.1.0 Bayes module.  ("Merlin Moncure", )

On Fri, 2005-07-29 at 09:37 -0800, Matthew Schumacher wrote:
> >
> > On my laptop this takes:
> >
> > real    1m33.758s
> > user    0m4.285s
> > sys     0m1.181s
> >
> > One interesting effect is the data in bayes_vars has a huge number of
> > updates and needs vacuum _frequently_.  After the run a vacuum full
> > compacts it down from 461 pages to 1 page.
> >
>
> I wonder why your laptop is so much faster.  My 2 min 30 sec test was
> done on a dual xeon with a LSI megaraid with 128MB cache and writeback
> caching turned on.

I only do development stuff on my laptop, and all of my databases are
reconstructable from copies, etc...  so I turn off fsync in this case.


> How often should this table be vacuumed, every 5 minutes?

I would be tempted to vacuum after each e-mail, in this case.


> Also, this test goes a bit faster with sync turned off, if mysql isn't
> using sync that would be why it's so much faster.  Anyone know what the
> default for mysql is?

It depends on your table type for MySQL.

For the data in question (i.e. bayes scoring) it would seem that not
much would be lost if you did have to restore your data from a day old
backup, so perhaps fsync=false is OK for this particular application.

Regards,
                    Andrew McMillan.

-------------------------------------------------------------------------
Andrew @ Catalyst .Net .NZ  Ltd,  PO Box 11-053, Manners St,  Wellington
WEB: http://catalyst.net.nz/            PHYS: Level 2, 150-154 Willis St
DDI: +64(4)803-2201      MOB: +64(272)DEBIAN      OFFICE: +64(4)499-2267
  What we wish, that we readily believe.
                                     -- Demosthenes
-------------------------------------------------------------------------


Attachment

pgsql-performance by date:

From: Karim Nassar
Date:
Subject: Re: Performance problems testing with Spamassassin
From: William Yu
Date:
Subject: Re: Performance problems on 4/8way Opteron (dualcore)