Norman Maurer wrote:
> No this will work well,
>
> how you did it ?
>
> bye
>
> On Mo, 2005-08-22 at 11:50 +0200, Massimo Lusetti wrote:
>
>> On 8/22/05, Norman Maurer <nm@spam-box.de> wrote:
>>
>>
>>> if you use the setup like in appliance you have to manually create each
>>> user in the dspam_virtual_uids table for which you want to scan email.
>>>
>> If i make my MTA (postfix) resolve the aliases before passing it (via
>> LMTP) to DSPAM i get the "final" real mailbox resolved and correctly
>> lookup up in the sql table.
>>
>> Is this a bad practice ?
>> Does this affect the efficency of DSPAM ?
>>
>> Regards
It improves the efficiency of sql lookups and other sql
functions(purge, re-index) and of dspam.
I'm doing that in my mta(qpsmtpd) by ldap and regex. Since
lmtp has only one dspam command as configured in dspam.conf
I will have to have the mta dspamc with other options for
dspam retrain aliases. That gets complicated when dspam
runs on a different pc(write to queue and use IPC:DirQueue
on both pc's and then a work process takes the message
out of queue and dspamc's it, or lmtp but not to dspam
and let that lmtp process on the other pc dspamc it).
I would like to see dspam use different dspam.conf
config options when running on different ports. Then
I could lmtp retrain as spam on port 1024, retrain as
innocent on port 2024. All lmtp, no dspamc.
-Bob
Received on Mon Aug 22 08:18:59 2005
This archive was generated by hypermail 2.1.8 : Thu Sep 29 2005 - 13:51:29 EDT