4 ms·

> bouncing increases the chance I'm becoming part of a back-scatter issue Bouncing doesn't cause backscatter, provided it's a real bounce; that means a bounce

> bouncing increases the chance I'm becoming part of a back-scatter issue Bouncing doesn't cause backscatter, provided it's a real bounce; that means a bounce that is a rejection by your mailserver, performed during the SMTP transaction. The message commonly known as a "bounce message" is technically known as a delivery status notification. The actual bounce message is composed by the sender's mailserver, and can't cause backscatter. Basically, a bounce sent after the SMTP server has said "Message accepted for delivery" is a bogus bounce, and causes backscatter. Sending bogus bounces is a reprehensible practice (and quite useless).

dspillett 5y ago
Aye. It sometimes happens because mail gets accepted then put into a local queue for processing (anti-spam checks and other scans) and the fake bounce sent after that. A badly designed system (and such do exist) will do this even for messages they could immediately reject based on the invalid incoming address. It can also happen because there is an intermediate SMTP server between the sender and receiver, the intermediate has dropped the connection to the sender before the final destination rejects, so if you want to bounce at all it has to be a fake bounce. I don't know without checking (and I've been too lazy to test!) if my mail services properly bounce based on target address or fake bounce, so I err on the side of not causing disruption if it is the bad one. I'll be replacing Zimbra with something else (probably self-built from standard parts as my needs are much more limited than when I picked Zimbra many years ago, just mail, nothing else) for my main mail service soon, this will be one of the things to make sure I get right.