Hi David,
I was able to disable checks defined in /var/lib/spamassassin by
making their score 0 in local.cf. That reduced the CPU usage a lot.
Rate-limiting isn't an option because my inbound servers receive mails
at a very high rate that incoming queue starts to increase
indefinitely.
But, after a day of figuring various options, I was finally able to
find some beefy boxes where I could run spamassassin and it's working
like a charm, exceeded my expectations.
Thanks
On Wed, Sep 18, 2013 at 4:40 AM, David B Funk
<dbf...@engineering.uiowa.edu> wrote:
> If you want to disable specific rules from the standard rules kit
> just set their score to zero in your local.cf config file.
> A rule with a score of zero isn't run.
>
> As the local.cf file is processed after the /var/lib/spamassassin contents
> that's how to over-ride the standard rules in a way that works regardless
> of updates.
>
> If you want to disable -all- rules, why are you even bothering to run
> sa-update?
>
> Note that CPU is still used even if all tests are disabled as SA needs to
> process the message to create the 'rawbody' 'full' and 'cooked' forms.
> The amount of CPU is dependent upon the size & content of the message
> (EG parsing a HTML message is more work than a plain text message).
>
> So if you're still hitting CPU load limits try limiting the message
> rates (at the MTA) and message sizes processed (skip large messages)
> or get a bigger proc.
>
>
>
> On Tue, 17 Sep 2013, Abhijeet Rastogi wrote:
>
>> Hi Dave,
>>
>> Thanks for the revelation. This clears a lot of things now. After I
>> disabled all the tests in 50_scores.cf, I was still seeing a lot of
>> CPU (a very heavily loaded server). This pretty much clears
>> everything.
>>
>> You guys have been of commendable help. One last thing, so is there a
>> way or a hack to disable rules under
>> /var/lib/spamassassin/3.003002/updates_spamassassin_org, in a way that
>> it survives sa-updates? Is there a patch in the wild that allows me to
>> do that?
>
>
> --
> Dave Funk University of Iowa
> <dbfunk (at) engineering.uiowa.edu> College of Engineering
> 319/335-5751 FAX: 319/384-0549 1256 Seamans Center
> Sys_admin/Postmaster/cell_admin Iowa City, IA 52242-1527
> #include <std_disclaimer.h>
> Better is not better, 'standard' is better. B{
--
Regards,
Abhijeet Rastogi (shadyabhi)
