Spam report addressed to multiple people

Glenn Steen glenn.steen at gmail.com
Tue Aug 19 14:28:57 IST 2008


2008/8/19 Andrew MacLachlan <andrew at gdcon.net>:
> Scott Silva wrote:
>>>>>
>>>>> But do you split messages before MailScanner? It sounds like you don't
>>>>> do that (properly).
>>>>
>>>> OK - any pointers to the _correct_ way of doing this with PF?
>>>>
>>> After spending some quality time with Google I found this:
>>>
>>> http://wiki.mailscanner.info/doku.php?id=documentation:configuration:mta:postfix:how_to:split_mails_per_recipient
>>>
>>> Is this still valid?
>>>
>>>
>> It should still be valid.
>>
> Does anyone have experience of using this configuration on a busy system? -
> I'm a little concerned about the caveats in the doc - what's the busiest
> system anyone would recommend this to be implemented on?
>
> -Andy
>
Which of the caveats concerns you most?
The concerns about performance isn't anything I've experienced, just
something logically less-than-optimal... The real hit in this would be
from SA, but if you use the SA results cache... that should mitigate
things pretty OK. I haven't got any substantial volume, so cannot say
how it will stand up to pressure on a "busy" server. It operates very
well with *our* distribution of multi-recipient mails and an average
of 6K messages/day on a rather weak desktop box... lab equipment that
during a time of ... crisis... had to step in and do the production
work (while I busily built a few new boxes:-).
As I say in the test (IIRC:-), I have very little use for it myself,
so ... the new boxes wasn't set up with it.

There are other effects, mainly with what you *cannot* do in a ruleset
while employing this setup, but I think those are covered in the
notes.

Cheers
-- 
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se


More information about the MailScanner mailing list