On Mon, Aug 07, 2006 at 07:43:38PM +0200, Fabien Tassin wrote: > Yes but with grey, the entry that triggered the flush will end-up in the db, right? > I mean, it seems the flush occurs *before* grey or auto-white are commited to the db. That should do it: acl greylist list "honeypots" delay 5d flushaddr Ok, it goes into the database, but with a delay so long it won't ever be used. And it will probably be flushed on the next spam that will fall into the spamtrap. > For now, I'm only experiencing various things to see what could work. > > In fact, i'm more interested by killing the whole flow so neither are good. > It's really auto-black that could do that. > I could still do a local dnsrbl fed by something monitoring the logs > for "black" events but I don't like spam interacting (write) with a DNS server > in production. You can have a local named bound to 127.0.0.1 that serves only that purpose. No need to send updates to your real domain. > I've done that in the past using nsupdate but I'd prefer > milter to do that internaly (less dependancies, no monitoring, etc). Yes, but the change is heavy, and making the code stable and reliable may take some time. Using a DNSRBL seems the most straightforward way. -- Emmanuel Dreyfus manu@...
Message
Re: [milter-greylist] blacklist
2006-08-07 by Emmanuel Dreyfus
Attachments
- No local attachments were found for this message.