I thought the idea behind sc_mailfilter.pl is that it searches headers for remote servers and logs them into the tarpit. What does no dirty headers mean?
I can't use a .forward file to get my spam to spamcannibal, but it seems I can use redirect. Trouble is, I don't know how to automate redirect. Seems refirect doesn't change the source address where normal forwarding does. As far as the comment that I'm running an abnormal email system, Postfix is commonly configured as an email gateway/relay host. If you have two relays like I do, it would be a problem to have mailboxes created on each one and not synchronized. The idea behind redundancy is not to create two email accounts but to ensure that there is more than one way for mail to come in. The obvious problem with relay hosts is that spamcannibal is expecting local delivery of spam on the mail relay. This is not realistic for the reasons already mentioned. One though is to move the relay smtp servers to port 26 and use a REDIRECT iptables rule, but the relay hosts are working where I will probably break them trying to change the port they run on. If I pull this off though, I can in theory run a local delivery enabled postfix server on port 25 and use fetchmail. What does sc_mailfilter.pl really do and what does it really need? Do I need to move my relay servers launched by MailScanner to port 26? If I process the spam on the central mailhub, what portions of spamcannibal do I need to install? I'm trying to protect my bandwidth or else I wouldn't need spamcannibal. As far as the mailhub having an outside connection, the only reason I do that is to ensure that I can email people. In theory, I should be able to route outbound email through my email gateways, but having a theory and putting it into practice are two different things. For one thing, I think you have to rewrite email headers to reflect the actual source when routing outbound email through a separate gateway. Because of the redundancy, I don't know what to rewrite the headers to until an email is going out. The easy solution is to let the mailhub send email out directly. Ideally though, the mailhub shouldn't have a direct Internet connection. Clearly, I could change from having three Internet connected servers to having just one. Trouble with this is, if I want to try something and potentially break email, there is no way around the break. With virtualization, one machine can pretend to be three. Trouble there is, I need to take some time to adapt to a single server over three seperate computers. I need to be able to split up network ports etcetera between virtual machines so that I can have machines that are Internet connected and ones that are not. Needless to say, switching to a single server infrasture will take time, it will take effort, and the switch will cost money. ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. _______________________________________________ PLUG mailing list [email protected] http://lists.pdxlinux.org/mailman/listinfo/plug
