[Raw Msg Headers][Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: calls to sleep() inside of smtpdata.c
Is it possible to make this an option for future versions of zmailer - just like the other tarpit options? I'm sure others may have experienced the same problem. We're going to make source modifications, to get around this - because a gateway in front of zmailer will cause the delay on valid traffic (relayed spam/valid mail).
Thanks,
-Richard
rhyatt@bluecatnetworks.com
Richard Hyatt
-----Original Message-----
From: Matti Aarnio [mailto:mea@nic.funet.fi]
Sent: Fri 8/8/2003 3:30 AM
To: Richard Hyatt
Cc: zmailer@nic.funet.fi
Subject: Re: calls to sleep() inside of smtpdata.c
On Thu, Aug 07, 2003 at 03:34:41PM -0400, Richard Hyatt wrote:
> What is the purpose of the calls to sleep(), which get exponentially
> longer if multiple messages come in on the same channel, when a message
> is sent to the freezer directory?
Huh ? Oh, that one.
> Our application uses the freezer directory to intercept messages that
> get quarantined into a separate database - the problem is that the
> exponential delay creates problems when the incoming connection is
> reused. We have remove the call to sleep() and things seem to work well,
> but would like to know why they were there in the first place - I
> wouldn't want to mess up some existing logic.
For some reason (not necessarily any very good one), the SMTP message
body receiver has a "tarpit" code of its own.
The basic idea has been (most likely) to slow down things, when a message
ends up in a freezer, but all in all I think these tarpits aren't very
good in deterring hostile spammers.
As spammers are either sending email in lots of separate connections,
or (which is easier to us), reusing connection, sleakest possible
server behaviour is at our (receiving side) interests.
> Thanks,
> -Richard
>
> Richard Hyatt
> rhyatt@bluecatnetworks.com
--
/Matti Aarnio <mea@nic.funet.fi>
{.n++%lzwmb맲r{9W{ay ʇڙ,jfhqz