>On Jun 29, 2015, at 12:35 PM, Reindl Harald <h.rei...@thelounge.net> wrote: >> >> Am 29.06.2015 um 18:29 schrieb Ted Mittelstaedt:
>What other free MTA is there that’s in common use? qmail is dead and >buried. Sendmail and Exim are pretty much niche. What exactly is wrong >with Postfix? Nothing. It's very powerful and fast with a lot of features built-in and plenty that can be added in added in as a milter. Postscreen is amazing with RBL weighting (just learned this recently from Reindl's suggestion). > I also do not undersrtand your (Ted) attitude, it really doesn’t seem >appropriate for this list. Chris, Ted, and others, Please don't take Reindl's comments as personal. He is always like this. Maybe English is not his first language or something and he doesn't understand how harsh he comes across. I have tried to email him off list to tone down his emails but nothing has changed. He has some good points and is very knowledgeable so we will have to take the bad with the good. I try to give him the benefit of the doubt that he simply means well. Email is not a place for emotion since there are no clues of body language or context so it's best to keep emails factual and free of opinion and expletives. >Absolutely. The amount of load you take off spamassassin by having the MTA >do some filtering (and it’s not an all-or-nothing affair with postscreen) is >pretty >amazing. I see no reason to let junk that far into the infrastructure when I >can >block it at the front door with so many fewer cpu cycles. Putting weighted >RBLs >up front also frees you up to re-examine your spamassassin config. And of >course you can toggle postscreen on a per-user basis if you have customers that >want all the spam… If you filter for large environments (more than a few hundred mailboxes) or don't use SA via a milter at SMTP-time, i.e. MailScanner, it's not possible to scan all of your email in SA and do it fast (a few seconds) without a lot of CPUs (across many servers). I block as much as possible in Postfix so the sender has proper feedback at the SMTP-level. And before everyone gets started on this being backscatter, it's not backscatter. Backscatter would be caused by accepting the message at SMTP-time then later bouncing it back to a potentially forged envelope-from address. Dave