Re: [Exim] exim v4, null envelope sender, 4xx from spamassas…

Pàgina inicial
Delete this message
Reply to this message
Autor: Dave C.
Data:  
A: Alan J. Flavell
CC: Exim users list
Assumpte: Re: [Exim] exim v4, null envelope sender, 4xx from spamassassin
On Sun, 25 Aug 2002, Alan J. Flavell wrote:

>
> (This may have an obvious general solution, but I haven't spotted it
> yet and we're quite new to using spamassassin).
>
> Consider this scenario: there's an incoming mail with a null
> envelope-sender, whose content is rated by SA to be in the twilight
> zone, where it issues a 4xx waiting for inspiration from the admin.
>
> The sending MTA responds by handing this off to our secondary
> MX. The secondary MX offers us the item again and, as it happens, the
> SA rating again falls into the twilight zone. The secondary MX then
> continues re-offering the item until it times out, unless we take some
> action.


You should really arrange for the scondary to issue the same responses
to the sender-smtp that your main server will. That is the 'obvious
general solution'


>
> I can't whitelist the sender because there isn't one. I can't apply a
> selective bonus (or malus) to the sending MTA, because the MTA is now
> our secondary MX (hi Chris). The item might be a bounce, or it might
> be a spam which is pretending to be a bounce: in a specific case, of
> course, I can view the SA-captured content and look at it - in fact in
> the particular instance which prompted this mail I _can_ see one way
> of killing it - but I can't see a measure which would be any
> more-generally applicable to future incidents. So I thought I would
> raise the more general question.
>
> Maybe, instead of returning 4xx to items which have a null envelope
> sender in this kind of situation, we might be better advised to accept
> and freeze them for inspection - and subsequent thaw or discard?
> Other thoughts, please?
>
> cheers
>
>
> --
>
> ## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##
>
>