Sysadmin wrote:
> I noticed some similar software (portgrey) are also capable to adjust
> the minimal count of mails before a client is whitelisted,
> milter-greylist accepts clients after one try, true or false?
True, and I think that's perfectly sufficient.
> And about memory and database, I compile milter-greylist with db4
> support, I understand, that all information are keeped memory and dumped
> for backup into db? But the db file is plain text based, so is it the
> database file at all?
I'm not aware that milter-greylist can be compiled with
db4 support. :-)
I'm using milter-greylist 2.1.1, it stores the greylist
in a plain text file.
> And does the dump file rotate ever?
No, it doesn't. Why should it? It is just replaced at
regular intervals (I set that to 10 minutes on my servers)
and when the process receives a SIGTERM.
> And how to check the actual memory use?
Use ps(1) or top(1).
There are quite a lot of greylist implementations. Let
me explain why I chose milter-greylist from all of them.
My primary MX is a Dual-Celeron-450 with 160 Mbyte RAM,
my secondary MX is a 486-DX2-66 with 32 Mbyte RAM. So
the hardware resources are quite limited.
Most of the greylist implementatios are written in Perl
and use an SQL backend (mostly MySQL). That's horribly
inefficient and completely out of the question.
I found two implementations written in C and without the
SQL DB requirement: The one from ACME:
http://www.acme.com/mail_filtering/
and the one from Emmanuel Dreyfus:
http://hcpnet.free.fr/milter-greylist/
The ACME implementation has only very few features and
doesn't even write anything to disk. So when you kill
the process, the greylist is lost. That's why I decided
to give Emmanuel's milter-greylist a try.
So far I'm quite satisfied, except for the the problem
I explained in a nother mail ("unknown users" being
greylisted), but that cannot easily be solved, so it's
not really the fault of milter-greylist.
Also, there seems to be a memory leak. Especially on
my 32Mbyte machine I see the milter process growing
quite fast, even though only a small portion of it
stays resident in memory. When I kill and restart it,
the process is smaller, even though it contains the
same greylist. My current work-around is to kill and
restart milter-greylist regularly via cron.
Best regards
Oliver
--
Oliver Fromme, secnetix GmbH & Co. KG, Marktplatz 29, 85567 Grafing
Dienstleistungen mit Schwerpunkt FreeBSD: http://www.secnetix.de/bsd
Any opinions expressed in this message may be personal to the author
and may not necessarily reflect the opinions of secnetix in any way.
We're sysadmins. To us, data is a protocol-overhead.