Protocol Error when using simscan

[Giovanni Bajo]

Hello,

we've been using simscan for a couple of years now. Lately, our mail

host mail.develer.com[83.149.158.210] said: 250 ok (in reply to DATA
command)

This problem happens once or twice a day in our very busy mail server. I
tried digging into qmail's source code trying to understand what might
cause this problem, but I get lost easily.

Can you think of *any* reason why this error could happen? My
understanding is that simscan is spawned as soon as the DATA command is
received... what if it hangs or exits abruptly for any reason? What if
it's killed by the kernel for any reason? Is there *any* way that such a
wrong reply to DATA command can be generated? Can you think of any?

Thanks!

[Matt Brookings]

Post by Giovanni Bajo
Hello,
we've been using simscan for a couple of years now. Lately, our mail

host mail.develer.com[83.149.158.210] said: 250 ok (in reply to DATA
command)

I'm not sure this is related to simscan. In fact, I would suggest that you are
not using qmail. However, see my comments below.

Post by Giovanni Bajo
This problem happens once or twice a day in our very busy mail server. I
tried digging into qmail's source code trying to understand what might
cause this problem, but I get lost easily.

This is exactly why I'm not sure you're using qmail. None of the message above
appears within the qmail source. You could have a very patched version of qmail,
of course.

Post by Giovanni Bajo
Can you think of *any* reason why this error could happen? My
understanding is that simscan is spawned as soon as the DATA command is
received... what if it hangs or exits abruptly for any reason? What if
it's killed by the kernel for any reason? Is there *any* way that such a
wrong reply to DATA command can be generated? Can you think of any?

None of your guesses above make sense. Once there is a reply to the DATA
command, the session can be terminated any way one likes, even if the RFCs
suggest you not do this.

My guess would be that your message contains a '.' (period) in a single line,
by itself, without be base64 or quoted-printable encoded, which causes a reply
from the receiving mail server too early.

My other guess would be, from my original assertion that you may not be using
qmail, is that your MTA has a huge bug in it.
- --
/*
Matt Brookings <matt-nNFyE46TO9nQT0dZR+***@public.gmane.org> GnuPG Key FAE0672C
Software developer Systems technician
Inter7 Internet Technologies, Inc. [phone]
*/