[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: OT: sorbs blacklisting scam



On Sun, Apr 30, 2006 at 04:16:17PM -0700, Mike Bird wrote:
> This thread has already mentioned numerous cases where
> backscatter is unavoidable.

Yes, there are some corner cases where email is sent automatically due to a
trigger. One good such example is registration confirmations for various
web-sites and mailing lists.

However, these are much more rare than the typical bounces such as "no such
user", or "mailbox full". It does not bother me if bounces are rare, and
surely almost all admins would agree that rare cases can be sorted out and
they are not a problem. But if a spammer can easily use lots of mail hosts
to reflect bounces containing their advertisements, that is bad. The more
they do that, the worse it is.

Abolishing bounces completely is not a reasonable target, but limiting
bounces only to necessary cases is a very important target.

The only excuse for "no such user"-bounces is short term configuration
mishap. It would be a bad idea to block hosts due to mailing list
confirmations or some other necessary and non-frequent automatic messages,
but if I recall the example presented in this thread mentioned a forged
sender causing the bounce. For that case, there is no excuse (not even if
Sorbs does it themselves aswell).

I am not going to give specific configuration examples, because lookup
configurations are always specific to the system and the specific needs of
it. Also there are many ways to copy account information to secondary
servers. The simplest case would be to copy the lookup-file by cron every n
minutes, a more complex way would be to setup replicated ldap servers and
make the MTA check accounts from there. But even that alternative is not
terribly complex compared to the task of running professionally an entire
mail system.

At least Exim4 has an excellent manual which includes ldap-lookups and many
other applicable types of lookups.

Attachment: signature.asc
Description: Digital signature


Reply to: