Will McDonald wrote:
> On 12/03/06, Sam Drinkard <sam@xxxxxxxxxx> wrote:
>
>> A while back, I posted a note asking if anyone had any ideas why the
>> /etc/mail/access file was not being parsed or utilized in the efforts to
>> stop spam and junk mail. I just looked over things again, and have still
>> not found any reason why it still permits the TLD's I have listed to pass
>> thru. I also thought perhaps there might be some "upper limit" to the
>> number of entries sendmail could handle. What do the sendmail guru's think
>> about that idea? I may reduce the number of entries from the current 275
>> +/- down to just the most offensive TLD's and see what happens. Short of
>> that, are there any other thoughts ya'll might have as to why it still
>> passes the stuff I want blocked?
>>
>
> I don't know the ins-and-outs of Sendmail access well but does it base
> its decision purely on the "From" address, which as we all know isn't
> necessarily where a message originates. Or could it be basing the
> access decision on the initial Received: from address, and/or that
> addresses reverse lookup, in the header?
>
> In which case, a spam could originate from mail.blah.com and access
> would accept it but the message itself would appear to come from
> spammers@xxxxxxxxxx You'd accept the message inspite of having .ru
> denied in your access.
>
> Just a thought.
>
> Will.
> _______________________________________________
> CentOS mailing list
> CentOS@xxxxxxxxxx
> http://lists.centos.org/mailman/listinfo/centos
>
>
>
As far as I know Will, sendmail looks at the access database, and will
not allow a connection from the sending host if that particular IP or
hostname happens to be in there. The access list *used* to work, but as
I mentioned, I'm wondering if perhaps I've hit an upper limit or
exceeded a limit where nothing in there is being parsed now. I don't go
by hostname when blocking. I look at the sending host IP and block
that. Headers from sendmail tell who or what connected to the port or
tried to connect.
Sam
