> Anders Norrbring wrote: > >> Anders Norrbring wrote: > >>> Is there a way to make Postfix send a '550 Recipient address > >> rejected' for > >>> every mail to a certain address in a domain, and still have it pass > >> the mail > >>> on to the IMAP server? > >>> I'm looking for a way to 'simulate' that the address does not > exist, > >>> although it does.. > >>> > >>> Anders > >>> > >> You need a smtpd_proxy_filter that responds with a 550 after > >> end-of-data and keeps a copy. For example, you can do this > >> with amavisd-new by telling it to reject with quarantine. > >> > >> Note that responding with 550 at end-of-data affects all > >> recipients of a multi-recipient message. > >> > >> -- > >> Noel Jones > > > > > > Thanks Noel, > > I'm just not sure about how to implement it.. I do use Amavis-New > already > > that I call from Postfix with 'content_filter = lmtp-amavis' in > main.cf > > > > Would you happen to have a concrete example on how to use it for > let's say > > '[EMAIL PROTECTED]' and '[EMAIL PROTECTED]' but not '[EMAIL PROTECTED]' ? > Or > > would it be possible to use only for a whole domain, and not some > addresses? > > In any case, it would be nice with an example! > > > > Thanks! > > Anders. > > > > > > Biggest change is you would need to use amavisd-new as a > smtpd_proxy_filter rather than a post queue content_filter. > Be aware that amavisd-new is not officially supported in this > mode, but is reported by others to work fine. Also note that > smtpd_proxy_filter is not recommended for high-volume sites > since the number of smtpd listeners you can run is limited to > the number of amavisd-new children your hardware can support. > > Then just blacklist that recipient, set amavisd-new to > D_REJECT, and turn on quarantine. > > I suggested amavisd-new as one example of how this could be > done. It's not necessarily the best way. > > -- > Noel Jones
Oh, okay... I have quite high volume, so I'm afraid that I can't take any chances by changing the way it works for doing this. The server is handling a number of other domains, so disturbing them with experimenting isn't really an option... Oh well, it's not a big deal anyway, but thanks for explaining! I'll look into it on an experimental setup at some point! Anders.
