[Kolab-devel] Kolab server - greylisting (long)
Gelpi Andrea
liste at gelpi.it
Tue Jan 22 12:47:26 CET 2008
Andreas Micklei wrote:
> Am Montag, 21. Januar 2008 schrieb Gelpi Andrea:
>> Andreas Micklei wrote:
>>> I have also installed postgrey and integrated it with kolab with a much
>>> simpler although not as clean approach:
>>>
>>> - install postgrey as delivered for your Linux distribution (yum install
>>> postgrey on CentOS)
>>> - configure postgrey to listen on a TCP Port (for example 60000,
>>> put 'OPTIONS="$OPTIONS --inet=60000"' into /etc/sysconfig/postgrey on
>>> CentOS) - put a check policy line with the correct port number into your
>>> kolab postfix main.cf ('check_policy_service inet:127.0.0.1:60000')
>>> - enable postgrey startup during system startup ('chkconfig postgrey on'
>>> on CentOS)
>>>
>>> In general just follow the instructions of your Linux distribution or
>>> general instructions to set up postgrey. Just put the
>>> check_policy_service line into the kolab postif main.cf instead of the
>>> one that comes with your distribution.
>> Please check postfix.sum I think now every mail is counted double.
>
> I read postfix.sum daily and it looks okayish to me. 76% was rejected
> yesterday due to greylisting, RBL or unknown receipient. Please explain what
> I should look for.
>
Well. Pflogsumm consider every connection to postfix as a new mail.
Using an inet socket instead of a unix socket postifx.sum count every mail more
than one time.
Look at postfix.log and you'll find a first connection from outside, than a new
connection from 127.0.0.1:1025 and another one from 127.0.0.1:1026.
Kolab use 2 inet socket (amavis and kolabfilter), so every mail is counted in
postfix.sum 3 times by pflogsumm script.
If you now add postgrey as an inet socket I think every mail is counted one more
time.
One year ago or perhaps more, I posted a modified pflogsumm script modified to
count every mail only one.
--
ing. Andrea Gelpi
***************************************************
La Terra non la abbiamo ereditata dai nostri avi,
ma la abbiamo presa in prestito dai nostri bambini.
***************************************************
More information about the devel
mailing list