Milter-greylist

4 \u0430\u0432\u0433\u0443\u0441\u0442\u0430 2016�\u0433. 20:43:34 CEST, "Carsten Grammes cgrammes@... [milter-greylist]" <milter-greylist@yahoogroups.com> \u043f\u0438\u0448\u0435\u0442:
>Am 04.08.2016 19:59, schrieb Mauricio Teixeira 
>mauricio.teixeira@... [milter-greylist]:
>> I started using DNSWL, then most of my problems with those senders 
>> that use multiple IPs have gone away.
>>
>That sounds a good workaround I surely will give a trial, thank you! 
>Nevertheless I would like to know whether milter-greylist has some 
>build-in feature to cope with multiple IPs (it should as it is spf
>aware).
>
>Carsten

Well, technically "it should" is wishful thinking ;)

In the end, it verifies if a specific remote IP address is a valid sender for a specific email domain. One address, as recorded in the database tuple (subject to later cidr-length matching as another poster reminded). It has lots of rules to let you check this, such as explicit IP addresses or cidr masks for hosts you know, previous automatic whitelisting, presence in dnsbl/dnswl services, or matching an spf record (and you should at least test it doesn't allow all internet including e.g. 127.0.0.1) among others.

So for general solution - there is none maybe. Reasonable ones were suggested; mixing spf-trust with a list of domains you know to have unspecified many relays and want to trust is a reasonable ruleset. DNSWL is also a good choice, make sure to place it before DNSBL as those often put a shadow on larger network ranges (e.g. whole ISPs one customer of which sent spam recently).

HTH,
Jim Klimov 
--
Typos courtesy of K-9 Mail on my Samsung Android