Hi Mella,
@Kloxo-DR: this would indeed be a problem if customers are using catchall for any of their domains as mentioned by @Chrisf and his reply on made me aware of this. So if you/your customers need to use catchall, you can't block non-existent recipients, i.e blocking prior to Qmail handling wouldn't make sense.
We are not intellectually poor to undertstand that catchall must not be catchall under certain circumstances! The idea of catchall is to catch all, regardless of if the recipient exists or not.
You cannot have a catchall for a domain for some recipients and not for others on the same domain. If catchall is activated, then Qmail shall receive all incoming emails.
Thus, it cannot make any recipient checks FOR THAT DOMAIN.
I am talking about the fundamental facility BEFORE CATCHALL is activated. If CHKUSER is properly compiled and configured, then one can use checking of a recipient, when catchall is deactivated.
The way how
checking of existent recipients from CHKUSER works is that the function
catchall is or could be deactivated and checking of recipient is activated. Emails to Non-existent recipients are then not accepted, denied or rejected.
The way how checking of non-existent recipients for CATCHALL works is that the function
catchall is activated and checking of recipient is deactivated. Emails to Non-existent recipients are then not accepted.
For this, one has to compile the Qmailtoaster to activate the function. Currently this seems not be the case. I have tried to activate it through the tcp.smtp rules but have failed.
So as you said, blocking prior to Qmail handling wouldn't make sense when you need catchall. This does make sense, when you need the opposite i.e. do not want catchall.
IF YOU DO NOT WANT CATCHALL WHY SHOULD YOU ACCEPT EMAILS? SIMPLY TO PROCESS AND DELETE THEM? Even that does not make sense, right?
Conclusion: Catchall and Recipient's check are contradictory functions. Thats where Chris made a mistake and claimed that he may not be able to use. Thats his mistake. He can peoperly use his qmail configuration further, for all his clients, if he requires catchall activated.
My proplem was that I did not need catchall. Although I had setup all accounts to delete, it did delete when I had sent test emails but did not delete when spammer sent thousands of emails per hour. Thats the reason why I needed to understand all this in a very time consuming manner.
I also see the confusion with everyone. Because this is certainly complex, I think it would be proper to explain everyone for a better understanding.
Is there any way of limiting/throttling sender/spammer? For instance limiting amount of connections within a time frame or amount of sent emails from the same sender within a time frame?
One must do this with connection tracking and having that port blocked if the number of connections per IP Address exceeds a number. This could be done at the level of a firewall.
Is Spamdyke configurable in a such way?
NO.
