You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Ramprasad <ra...@netcore.co.in> on 2006/11/18 09:59:57 UTC

Bayes file or SQL

Which is lighter to use bayes in files or bayes in SQL for a large setup
We get around 6-7 Million mails per day on our 14 servers  ( >80% get
rejected at MTA however)

Currently each machine( dual Xeon 4GB Ram) running Postfix + SA +
MailScanner has it own bayes files. How about running a mysql server
with common bayes 

Thanks
Ram



Re: Bayes file or SQL

Posted by Matt Kettler <mk...@verizon.net>.
Ramprasad wrote:
> On Sat, 2006-11-18 at 07:29 -0500, Matt Kettler wrote:
>   
>> Ramprasad wrote:
>>     
>>> Which is lighter to use bayes in files or bayes in SQL for a large setup
>>> We get around 6-7 Million mails per day on our 14 servers  ( >80% get
>>> rejected at MTA however)
>>>   
>>>       
>> For such a large system, SQL would be by far more efficient.
>>
>> Admittedly SQL is "heavier" than bayes in terms of memory overhead,
>> because you need to run a SQL daemon. However, the performance gain is
>> substantial.
>>
>> See  http://wiki.apache.org/spamassassin/BayesBenchmarkResults
>>     
>>> Currently each machine( dual Xeon 4GB Ram) running Postfix + SA +
>>> MailScanner has it own bayes files. How about running a mysql server
>>> with common bayes 
>>>       
>
> SDBM seems much better than anything else. 
> So is that not the best to use 
>
>
>   
I personally use SDBM. However, I'll warn you that very few use SDBM.
When I converted, I found bugs in the sa-learn --import process when
using SDBM. I was able to work around them, and I think they've been
fixed since, but be aware that SDBM is not very commonly used, nor is it
well tested.

That said, I've been happy with it, but my server is small-scale.


Re: Bayes file or SQL

Posted by Ramprasad <ra...@netcore.co.in>.
On Sat, 2006-11-18 at 07:29 -0500, Matt Kettler wrote:
> Ramprasad wrote:
> > Which is lighter to use bayes in files or bayes in SQL for a large setup
> > We get around 6-7 Million mails per day on our 14 servers  ( >80% get
> > rejected at MTA however)
> >   
> For such a large system, SQL would be by far more efficient.
> 
> Admittedly SQL is "heavier" than bayes in terms of memory overhead,
> because you need to run a SQL daemon. However, the performance gain is
> substantial.
> 
> See  http://wiki.apache.org/spamassassin/BayesBenchmarkResults
> > Currently each machine( dual Xeon 4GB Ram) running Postfix + SA +
> > MailScanner has it own bayes files. How about running a mysql server
> > with common bayes 
> 

SDBM seems much better than anything else. 
So is that not the best to use 


Thanks
Ram



Re: Bayes file or SQL

Posted by Matt Kettler <mk...@verizon.net>.
Ramprasad wrote:
> Which is lighter to use bayes in files or bayes in SQL for a large setup
> We get around 6-7 Million mails per day on our 14 servers  ( >80% get
> rejected at MTA however)
>   
For such a large system, SQL would be by far more efficient.

Admittedly SQL is "heavier" than bayes in terms of memory overhead,
because you need to run a SQL daemon. However, the performance gain is
substantial.

See  http://wiki.apache.org/spamassassin/BayesBenchmarkResults
> Currently each machine( dual Xeon 4GB Ram) running Postfix + SA +
> MailScanner has it own bayes files. How about running a mysql server
> with common bayes 


Re: Bayes file or SQL

Posted by Martin Hepworth <ma...@solidstatelogic.com>.
Ramprasad wrote:
> Which is lighter to use bayes in files or bayes in SQL for a large setup
> We get around 6-7 Million mails per day on our 14 servers  ( >80% get
> rejected at MTA however)
> 
> Currently each machine( dual Xeon 4GB Ram) running Postfix + SA +
> MailScanner has it own bayes files. How about running a mysql server
> with common bayes 
> 
> Thanks
> Ram
> 
> 
g'day

you'll prob get more performance out of SQL for that load, and of course 
benefit from using the same bayes data on all you machines

-- 
Martin Hepworth
Senior Systems Administrator
Solid State Logic
Tel: +44 (0)1865 842300

**********************************************************************

This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.

This footnote confirms that this email message has been swept
for the presence of computer viruses and is believed to be clean.	

**********************************************************************