SMTP-AUTH mail being marked as spam

[Valmiki N. Ramsewak]

On Apr 27, 2008, at 6:32 PM, Gerard wrote:

> On Sun, 27 Apr 2008 18:30:27 +0100
> Julian Field <MailScanner at ecs.soton.ac.uk> wrote:
>
>>
>>
>> Gerard wrote:
>>> On Sun, 27 Apr 2008 17:14:53 +0100
>>> Julian Field <MailScanner at ecs.soton.ac.uk> wrote:
>>>
>>>
>>>> Valmiki N. Ramsewak wrote:
>>>>
>>>>> Hi,
>>>>> 	I run the latest version of mailscanner and postfix 2.3.6.
>>>>> I send mail from my laptop, via the mail server. I have SMTP-AUTH
>>>>> enabled so only credentialed users can send. However any mail sent
>>>>> this way is being marked as spam. This happened to me before and I
>>>>> realized the problem was I wasn't having postfix add the
>>>>> (authenticted user = xyz) line in the mail headers, so I included
>>>>> that and it worked. Now I'm not sure what the problem is.
>>>>>
>>>>> 	If I login to my my mail server and send mail from mutt it
>>>>> works just fine and doesn't mark it as spam.
>>>>>
>>>>> I do want mailscanner to scan the mail, just not mark
>>>>> authenticated mail as spam, but check for viruses.
>>>>>
>>>>> The spamassassin score is -ve.. but it says its being marked as
>>>>> spam because of spamhaus-ZEN. Granted I'm on a DSL connection with
>>>>> dynamic ip, so I have no control over what my ip address is when I
>>>>> reconnect.
>>>>>
>>>> You can't use a blacklist that includes dial-up dynamic addresses
>>>> on your server, connect from a dial-up dynamic address, and not
>>>> expect to get blacklisted. Simple logic :-)
>>>>
>>>> Can you move your blacklist checking into your MTA and have Postfix
>>>> not apply blacklist checks to authenticated SMTP connections?
>>>>
>>>> By the time MailScanner gets at it, it doesn't know whether you
>>>> were authenticated or not (but you could write a simple Custom
>>>> Function to set the "Spam List" setting to different values
>>>> depending on the first header in the message, and look for the
>>>> signs that your mail server thinks you are authenticated).
>>>>
>>>> Jules
>>>>
>>>
>>> Are you the only user on the system? If so, just send via port 587.
>>> Do not allow Mailscanner to touch that port. You will still receive
>>> mail via port 25, which is scanned by Mailscanner. It is fairly
>>> trivial to set that up on a Postfix system.
>>>
>> If you do that, you won't get any virus scanning either, which is
>> rather a dangerous thing to do.
>
> If he is the only user of the system, there really is not any risk. He
> should know whether he is sending SPAM or not. I know several  
> instances
> of Postfix configured in exactly that manner. Now, if he is hosting a
> mail server for others, then that is a different matter. I was not  
> lead
> to believe that however. He still enjoys the protection of Mainscanner
> on his inbound mail (port 25) so I really do not see the problem.
>
> Just my 2¢.


Just to clarify there are other users/domains on my server.

But while theoretically your solution is "safe" for a single user, for  
windows users there is too much risk if the system is exploited (you'd  
think someone capable of running a mail server can keep a win machine  
updated but you never know).  I like the idea of having all entrances  
guarded, not having a secret entrance unguarded with only a secret to  
get in.

My setup is  port 25 is regular, only sends/relays from localhost..  
port xx is widely open to any network for sending/relaying, but you  
must authenticate or your mail will not go through.  They all end up  
in mailscanner after that :)