Re: Backscatter Mail
- Actually, the program a) grabs pop3 object b) parses it c) modifies
subject slightly with certain keywords based on email contents d)
based on it builds smtp object. Hence, the from and to of the smtp
object is being assigned by from and to of the parsed pop object. So,
it is not a resend. The to address is always email from our domain
which is hosted by yahoo.
The reason why we need to send to the same from and to addresses is
for agents to process email based on an outlook folder determined by
subject rule. Also, these agents need the ability to respond to these
As I am writing to you, I am thinking a) I should create mail store
locally for the same domain. b) create mx record for the domain only
on the local(ie no outside box is using this dns server) postfix
server c) point deliveries via lmtp to mail store.
Please, let me know if this sounds right.
I didn't plan it. I inherited it. Though there are very few of these
occurrences, I am concerned now that we may get blacklisted for
something like this.
On Wed, Jul 31, 2013 at 9:38 PM, Viktor Dukhovni
> On Wed, Jul 31, 2013 at 09:31:20PM -0400, Roman Gelfand wrote:
>> We are running an application which pulls mail from yahoo pop3 server
>> and ultimately resends it, via local postfix server, keeping the same
>> from and to addresses.
> This is not generally possible. POP3 does not record the envelope
> addresses, and resending mail to the "To:" and "Cc:" addresses of
> already delivered mail is a disaster in the making. Duplicate
> deliveries, mail loops, ...
> The only thing that is safe to do is to deliver to a fixed local
> address that is associated with the remote mailbox.
> Can you explain how you determine which addresses to redeliver to?
>> The way I see it, there are rare instances
>> where from domain requires strict dkim enforcement. As a result,
>> yahoo rejects it notifying this postfix server.
> DKIM signs content. If you don't modify the message, its DKIM
> signature remains valid.
> It sure sounds like you're doing something unusual or even unwise.
> What real problem are you solving?