Question about sa-learn --no-sync

[Steve Bergman <sb...@gmail.com>]

Hi,

I just have a question about the expected performance of salearn with 
the --nosync option. Working with the default backend, with a bayes_toks 
file size of less and 1 MB, it's taking about 1.5 seconds to learn 1 
small email message without --no-sync, and about 1.3 seconds with 
--no-sync. Are these times reasonable? I had read in the docs that 
--no-sync was supposed to speed things up greatly.

Learning large numbers of emails with a single sa-learn command is quite 
fast. It can learn 1000 emails in about 15 seconds. But I have need to 
do a single mail at a time as quickly as possible, with a sync happening 
later.

This is with SA 3.3 as shipped with the latest Debian Stable.

Thanks for any feedback,
Steve Bergman

Re: Question about sa-learn --no-sync

[Steve Bergman <sb...@gmail.com>]

On 06/22/2014 05:14 AM, Paul Stead wrote:
> Hi Steve,
>
> If you run sa-learn with the -D debug option, you'll see that most of 
> the time is during startup.

Yeah, I kinda figured. But I'm sure I read in the wiki or somewhere that 
--no-sync was supposed to be "much faster" for learning individual emails.

>
> I believe - correct me if I'm wrong - that spamc is the better option 
> to use as it will connect to existing spamd processes and is much 
> quicker overall
>

Thank you *so* much. It's lightning fast. It took me from 1.5 seconds 
per message moved into the Junk folder to 7 messages per second. And 
that's with dovecot-antispam's overhead included.

-Steve

Re: Question about sa-learn --no-sync

[Paul Stead <pa...@zeninternet.co.uk>]

Hi Steve,

If you run sa-learn with the -D debug option, you'll see that most of the time is during startup.

I believe - correct me if I'm wrong - that spamc is the better option to use as it will connect to existing spamd processes and is much quicker overall

Check it out - http://spamassassin.apache.org/full/3.3.x/doc/spamc.html

Paul

On 22/06/14 02:15, Steve Bergman wrote:
Hi,

I just have a question about the expected performance of salearn with the --nosync option. Working with the default backend, with a bayes_toks file size of less and 1 MB, it's taking about 1.5 seconds to learn 1 small email message without --no-sync, and about 1.3 seconds with --no-sync. Are these times reasonable? I had read in the docs that --no-sync was supposed to speed things up greatly.

Learning large numbers of emails with a single sa-learn command is quite fast. It can learn 1000 emails in about 15 seconds. But I have need to do a single mail at a time as quickly as possible, with a sync happening later.

This is with SA 3.3 as shipped with the latest Debian Stable.

Thanks for any feedback,
Steve Bergman

--
Paul Stead, Zen Internet
Systems Engineer
Direct: 01706 902018
Fax: 01706 902170
Web: zen.co.uk

Fibre Optic Broadband offers download speeds as high as 76Mbps. Check your Fibre broadband availability here: http://www.zen.co.uk/business/broadband/fibre-optic-broadband.aspx

This message is private and confidential. If you have received this message in error, please notify us and remove it from your system.

Zen Internet Limited may monitor email traffic data to manage billing, to handle customer enquiries and for the prevention and detection of fraud. We may also monitor the content of emails sent to and/or from Zen Internet Limited for the purposes of security, staff training and to monitor quality of service.

Zen Internet Limited is registered in England and Wales, Sandbrook Park, Sandbrook Way, Rochdale, OL11 1RY Company No. 03101568 VAT Reg No. 686 0495 01